text stringlengths 8 5.77M |
|---|
Introduction {#Sec1}
============
Atherosclerosis is a major contributor to global morbidity and mortality; cardiovascular disease was responsible for 17.5 million deaths in 2012^[@CR1]^. Addressing this health challenge requires lifestyle changes, development of potent anti-atherosclerotic drugs, in addition to more accurate diagnostic tools for early detection. Atherosclerosis is a complex process, with many steps including, accumulation and oxidation of LDL, infiltration of inflammatory cells, and smooth muscle cell migration from the medial to the intimal layer of the vessel wall^[@CR2],\ [@CR3]^. Nevertheless, lipoprotein (LP) deposition in the subendothelial space of the blood vessel represents the initial step for the development of cardiovascular diseases. Due to the irreversibility of the disease, and the role played by lipoprotein deposition, attempts have been made to monitor lipoprotein binding to endothelial-like surfaces under accelerated conditions^[@CR4]^. This has then been correlated to various other biomarkers, as well as to clinical results, e.g., in studies on type-2 diabetes, coronary by-pass, and metabolic syndrome patients^[@CR4],\ [@CR5]^. However, the basic mechanisms for lipoprotein deposition at plasma membrane surfaces are still not fully understood despite the important role of LP deposition on the formation of foam cells and subsequent atherosclerotic plaques.
HDL and LDL are macromolecular assemblies of lipophilic cholesterol esters and triglycerides, encapsulated by a phospholipid monolayer and apolipoproteins^[@CR6]^. These two lipoprotein classes (HDL and LDL) differ in density, apolipoprotein type, lipid content and composition^[@CR7]^. HDL and LDL are often referred to as 'good' and 'bad' cholesterol due to their correlation with the development of atherosclerosis being negative and positive, respectively^[@CR8]^. Such correlation between lipoprotein class and clinical outcome is well known: LDL-derived cholesterol accumulates in the vessel wall during atherosclerosis, whereas HDL is believed to facilitate removal of cholesterol from lipid-loaded foam cells present in the vascular wall, the latter being referred to as reverse cholesterol transport^[@CR9]^. However, there is an urgent need for an improved understanding of the factors controlling lipoprotein interactions with cell membranes and membrane components. This is of direct importance for the investigation of the formation of foam cells and subsequent atherosclerotic plaques, and for the determination of better clinical markers for atherosclerosis. For example, lipid exchange (where no net transfer takes place) and lipid transfer between lipoproteins of different types (i.e. between HDL and LDL) are known to occur, both *in vitro* and *in vivo*, for phospholipids, cholesterol, and sphingomyelin^[@CR10]--[@CR13]^. Other studies found that lipid exchange is independent of protein exchange^[@CR14]--[@CR17]^. However, quantification of lipid exchange and transfer is challenging and examples are scarce in literature. Previous studies primarily concentrated on exchange between lipoprotein particles, although a few studies were performed to follow the exchange between lipoproteins and lipid microemulsions^[@CR18],\ [@CR19]^, lipid vesicles^[@CR20]^, and cells^[@CR21]^. These studies used chemical analysis and focused on net changes in lipoprotein particle composition upon a given equilibration time. Therefore, these studies did not monitor the exchange processes directly. For quantitative studies, well-defined samples are required in order to control and measure, with time, the relative proportions of lipoproteins both in terms of the number of particles and the lipid concentration available for exchange.
Supported lipid bilayers (SLBs) are commonly used as models of cell membranes, as they allow the study of many processes *in situ* using highly surface sensitive techniques such as quartz crystal microbalance with dissipation (QCM-D^[@CR22]^) and neutron reflection (NR^[@CR23]^). QCM-D measures wet mass changes at the surface down to a few ng/cm^2^ ^[@CR24]^, while NR provides information on the structure of complex, buried interfaces down to a few Å resolution in the direction perpendicular to the interface (see for example ref. [@CR25]). In this work, we first study the timescale of interaction between the lipoproteins and SLBs by following lipoprotein adsorption using QCM-D. Then, we employ both hydrogenated (non-deuterated) and tail-deuterated phospholipids to study time-resolved interactions between lipoproteins and SLBs using NR with contrast variation^[@CR25]^.
In order to demonstrate the power of NR to decouple lipid transfer (and other components such as proteins) from lipid exchange between the lipoprotein and the bilayer, simulated reflectivity profiles are given in Fig. [1](#Fig1){ref-type="fig"}. A 4-layer model consisting of head group- tail region- head group with a water layer between the bilayer and the surface, as often seen in NR of SLBs was used to describe the lipid bilayer at the surface (SI Fig. [S1](#MOESM1){ref-type="media"})^[@CR26]^. The profiles in Fig. [1](#Fig1){ref-type="fig"} show a hydrogenated (h-) and tail-deuterated (d-) SLB in D~2~O and H~2~O subjected to different degrees of lipid removal from the SLB and lipid exchange with hydrogenated material. For the h-SLB, removal of lipids by the lipoproteins will be very evident in the D~2~O contrast (Fig. [1a](#Fig1){ref-type="fig"}), as the volume previously occupied by a hydrogenated lipid is replaced by deuterated water of higher scattering length density (SLD). However, removal of the h-SLB in H~2~O will be masked as the water that replaces the hydrogenated lipids has a similar SLD (Fig. [1b](#Fig1){ref-type="fig"}). Any deposition of hydrogenated material from the lipoprotein to the hydrogenated bilayer will not be evident in H~2~O and D~2~O (not shown) due to the lack of contrast between the original lipids and deposited material. For the d-SLB, on the other hand, NR is sensitive to both removal and exchange mechanisms as shown in Fig. [1c--f](#Fig1){ref-type="fig"}. The removal (Fig. [1c](#Fig1){ref-type="fig"}) and exchange (Fig. [1e](#Fig1){ref-type="fig"}) of deuterated molecules for hydrogenated ones can be clearly discerned in the D~2~O contrast, as the SLD of the whole SLB or just the tail region is lowered due to increased solvent penetration or incorporation of hydrogenated material, respectively. It is, however, not possible to completely decouple the exchange and removal processes when measuring a deuterated bilayer in H~2~O (Fig. [1d and f](#Fig1){ref-type="fig"}); in both cases the volume occupied by the deuterated molecule is replaced by hydrogenated material, either water or deposited hydrogenous molecules from the lipoprotein. This contrast is very sensitive to the overall change in the SLD of the bilayer and is thus used to track the kinetics of the system in this study. Taken together, these considerations demonstrate that quantitative monitoring of exchange and transfer processes between lipoproteins and the membrane by NR requires the use of multiple isotropic contrasts: Not only the use of heavy and light water are needed but also deuterated and hydrogenated lipids are required. In contrast to previous exchange studies, which relied on transfer of ^14^C-, ^32^P- or ^3^H-labeled molecules or serum and did not allow for *in situ* measurements^[@CR10],\ [@CR14]--[@CR17],\ [@CR21],\ [@CR27]^, the present approach enables studying non-labelled HDL and LDL extracted from human serum in environments close to those found in the body, and monitoring the exchange *in situ* with lipid bilayers.Figure 1Simulated reflectivity profiles for hydrogenated (h-) or tail deuterated (d-) SLBs. (**a**) and (**b**) show the effect of lipid removal from the bilayer on a h-SLB in D~2~O and H~2~O, respectively, while (**c**,**d**) show the removal process on a d-SLB. (**e**,**f**) show the effect of replacing tail-deuterated lipid molecules in the d-SLB with hydrogenated material from a model lipoprotein in both D~2~O (**e**) and H~2~O (**f**). Inset to each graph depicts lipid exchange or removal described in the simulated reflectivity changes. SLD is shown on a greyscale with material of a high SLD (deuterated) in black and low SLD (hydrogenated) in white. For clarity rearrangement after lipid removal has not been depicted, further discussion can be found in the text.
In summary, we present an alternative approach to monitor *in situ* lipid transfer between lipoprotein particles and model cellular membranes (SLBs) of known composition under controllable, reproducible conditions in a non-destructive manner. As a proof of concept, we use SLBs composed of hydrogenated and tail deuterated 1,2-dimyristoyl-*sn*-glycero-3-phosphocholine (DMPC) with 10 mol% 1,2-dimyristoyl-*sn*-glycero-3-phospho-L-serine (DMPS) to impart a slight negative charge close to the zeta potential of natural cell membranes^[@CR28],\ [@CR29]^. The head groups of these lipids occur in endothelial cell membranes, although it is mainly present in the inner leaflet of healthy cells^[@CR28]^. First, we study the timescale of lipoprotein (purified from healthy human patients) adsorption to SLB by quartz crystal microbalance with dissipation (QCM-D). Then, we employ both hydrogenated and tail-deuterated phospholipids to follow the time-resolved interactions between lipoproteins and SLBs using neutron reflection (NR) with contrast variation. By knowing the lipoprotein particle concentration (content of lipid and protein) and controlling the SLB composition, this method allows lipid transfer and lipid exchange to be studied in detail and can now be applied to follow a multitude of factors thought to affect the risk for atherosclerosis. These include for example, co-incubation with HDL and LDL at various lipoprotein ratios, the type of lipids in the membrane (saturated fats and cholesterol), genetic disposition or environmental factors^[@CR7]^.
Results and Discussion {#Sec2}
======================
Kinetics of lipoprotein deposition on SLB {#Sec3}
-----------------------------------------
Figure [2](#Fig2){ref-type="fig"} shows the relative change in frequency (Δf, negative values) and dissipation (ΔD, positive values) for QCM-D signals upon introduction of a lipoprotein solution to a preformed SLB containing 90 mol% DMPC and 10 mol% DMPS. Data obtained from the preparation of the SLB on the silicon sensor is provided in Fig. [S2](#MOESM1){ref-type="media"} in the Supplementary Information (SI). Lipoproteins were purified from human serum blood using sequential ultracentrifugation and size exclusion chromatography (SI Fig. [S3](#MOESM1){ref-type="media"}). After lipoprotein introduction, there was a lag-time of approximately 45 minutes before lipoprotein adsorption was sensed. Then, very slow adsorption kinetics took place, with no steady state reached during the time frame of the experiment (12 hours). It can be clearly seen that adsorption of LDL (dark colours) caused a larger Δf and ΔD response compared to HDL (pale colours). Furthermore, the overtones showed a greater degree of spreading for LDL over time.Figure 2QCM-D data of a supported DMPC/DMPS (90:10 mol%) bilayer upon exposure to HDL (pale) or LDL (dark). Changes in frequency and dissipation are shown in blue and red color, respectively. Three overtones are plotted: 7^th^ (solid line), 9^th^ (dashed line) and 11^th^ (dotted line).
Overtone spreading often indicates the presence of an diffuse and heterogeneous adsorbed layer^[@CR30]^. The difference between the two lipoprotein responses could be due to: (1) different extent of adsorption and (2) differences in lipoprotein size (HDL and LDL radii are roughly 5 and 12 nm respectively, see SI Figure [S2](#MOESM1){ref-type="media"}) and rigidity. The greater signal observed for LDL arises not only from a higher lipoprotein adsorption density but also from hydrodynamic contributions, where a larger number of water molecules oscillate with the quartz crystal for thicker layers at partial coverage^[@CR30]^.
Although QCM-D is a simple and sensitive technique for studying the adsorbed wet mass of lipoproteins to lipid bilayers, it does not allow detailed structural analysis and studies of exchange and uptake from the SLB. Fortunately, however, the slow adsorption of lipoprotein to the surface (equilibrium not reached at 12 hours) is ideal for studies using neutron scattering methods, as the change can be followed over time, with repeated full Q-range characterisation, without a large change in the reflectivity during each single measurement. In order to quantify the effect of lipoprotein addition on the SLB structure and composition by NR, the SLB was characterised prior and after lipoprotein adsorption and the results of the modelling analysis is described in the following sections.
Structural characterisation of the SLB prior to lipoprotein addition {#Sec4}
--------------------------------------------------------------------
The NR data for the SLBs formed (via vesicle fusion^[@CR22]^) using h- and d-lipids were simultaneously fitted to a single model. The best fit to the NR data (Table [1](#Tab1){ref-type="table"}, fits shown in SI Figure [S4](#MOESM1){ref-type="media"}) was obtained using the model described above and in Supplementary Information [S1](#MOESM1){ref-type="media"}. The model was further constrained by calculation of the mean molecular area (MMA), such that the head and tail regions of each molecule must occupy the same proportion of the layer. In all bilayers a MMA of 58 ± 1 Å^2^ was obtained, in good agreement with the values reported in literature^[@CR31]^. The bilayers could also be well fitted to literature values for thickness and hydration of head and tail regions and displayed complete surface coverage^[@CR31]^. A better fit to the data was found if the head group regions were thinner and less hydrated than in pure PC SLBs, likely due to some tilting of the head group, as reported previously for SLBs composed of charged lipids on silicon surfaces^[@CR32]^.Table 1Fitted SLD values of components of the lipid bilayer system and physical parameters obtained from fitting the 90:10 mol% DMPC/DMPS SLB prior to lipoprotein exposure. Typical errors are 1 Å for thickness, and 0.05 for volume fraction.MaterialSLD (\*10^−6^ Å^−2^)Thickness (Å)Volume FractionSilicon2.07---Silicon Oxide^a^3.4780.98^a^Head groups in H~2~O^b^1.896.50.9Head groups in CMSi^b^1.93Head groups in D~2~O^b^1.99d-Tail6.7^c^26.51.0h-Tail−0.3^a^One silicon block was found to have a silicon oxide volume fraction of 0.83. ^b^Calculated assuming 10 mol% PS heads, with exchangeable hydrogens. ^c^Small deviation from the theoretical value for the SLD of d-tail (6.89 × 10^−6^ Å^−2^) due to incomplete lipid deuteration, as confirmed by NMR spectroscopy (SI Fig. [S4](#MOESM1){ref-type="media"}).
Kinetics for change in lipid composition of the SLB depends on lipoprotein class {#Sec5}
--------------------------------------------------------------------------------
HDL or LDL, at fixed concentration in terms of the number of lipoprotein particles, in hTris (Tris buffer in H~2~O) was incubated for 8 h on both h- and d-SLBs, and characterised over time. Kinetics for lipoprotein interaction were measured every 10 min during the first hour (SI Fig. [S6](#MOESM1){ref-type="media"}) and then every two hours during 8 h (Fig. [3a,b](#Fig3){ref-type="fig"}). In the d-SLB experiment, a reduction of the reflectivity indicates removal of deuterated lipids from the bilayer. It is not possible to distinguish between removal and exchange of lipids using the hTris contrast (see Fig. [1](#Fig1){ref-type="fig"}), so in modelling this data we assumed only lipid removal occurred and used this value to calculate a percentage of deuterated bilayer removed either by exchange or actual lipid removal (Fig. [3c](#Fig3){ref-type="fig"}). Using this contrast for the measurements during incubation, however, circumvents any artefacts in the kinetics of interaction, which can arise from the use of a deuterated buffer; an effect that has been previously discussed for other systems^[@CR33],\ [@CR34]^. The kinetic data was fitted allowing for small structural changes in the bilayer with no adsorbed lipoprotein layer on top. The difference in SLD between hTris and the lipoprotein is low and this contrast is thus insensitive to such an adsorbed layer, especially for low coverage. Figure [3c](#Fig3){ref-type="fig"} shows the percentage of lipids removed from the bilayer over time. Importantly, it can be seen that HDL removes more lipids from the bilayer compared to LDL. The removal of lipids is linear with respect to time, suggesting that no saturation was reached within the time frame of the experiment. No significant change in the reflectivity of h-SLB profiles could be detected over time, except in the high Q data. This could be related to a decoupling of the head group regions by roughening and removal of lipids (SI Fig. [S7](#MOESM1){ref-type="media"}).Figure 3Fitted (lines) experimental (markers) data for a 90:10 mol% DMPC/DMPS d-SLB incubated with (**a**) HDL and (**b**) LDL over 8 hours in hTris. (**c**) Percent of lipid removed from dSLB after incubation with HDL (red markers) or LDL (blue markers) over time calculated from best fit assuming lipid removal.
Exchange of lipids either between the two leaflets of a bilayer or between the bilayers of separate vesicles was previously studied experimentally using neutron scattering^[@CR35]--[@CR37]^ and other techniques^[@CR38]--[@CR41]^. For example, total exchange of lipids between a d-SLB and excess h-vesicles was found to occur exponentially over 11 hours^[@CR35]^. This is very different to our data on lipoproteins, where a linear change is observed over 8 hours. The loosely packed nature of the lipid monolayer in the lipoproteins, and fast translational diffusion of lipids at the lipoprotein surface may aid the organization and exchange of the lipoprotein cargo within the particle and with the cell wall thereby affecting the observed kinetics^[@CR42]--[@CR44]^. Moreover, no significant lipid exchange or removal seems to occur during the first hour of lipoprotein incubation (SI Fig. [S6](#MOESM1){ref-type="media"}) in agreement with the lag time for lipoprotein adsorption observed by QCM-D (Fig. [2](#Fig2){ref-type="fig"}). This suggests that lipoprotein adsorption to a SLB is a pre-requisite for any lipid exchange to occur. Exchange therefore takes place upon direct contact between the lipoproteins and the SLB. Thus, passive diffusion through the solution has a negligible effect due to the low solubility of phospholipid molecules in aqueous solution and therefore it is likely that only the lipids in the lipoproteins bound to the SLB are able to exchange with the SLB. The presence of specific lipoprotein receptors in the membrane may reduce the lag time observed for lipoprotein binding to the membrane and thus accelerate lipid transfer in this way.
Quantification of lipid removal and exchange by lipoproteins {#Sec6}
------------------------------------------------------------
For quantification of the SLB composition and determination of its structure after lipoprotein addition, NR measurements were taken after rinsing in 3 contrasts (dTris, hTris and a mixture of h/dTris that matches SiO~2~ (CMSi)) and no significant changes in reflectivity were seen due to rinsing, suggesting that either adsorbed lipoprotein particles were well adhered to the bilayer, the coverage was too low or there was insufficient contrast between the LP and hTris to be detected in the hTris contrast (SI Fig. [S8](#MOESM1){ref-type="media"}). Figure [4](#Fig4){ref-type="fig"} shows the NR data and best fits for both h- and d-SLBs after HDL or LDL in three isotropic contrasts. Table [2](#Tab2){ref-type="table"} summarises the main structural information extracted from the modelling while the corresponding SLD profiles calculated from the fits can be found in the Supplementary Information (SI Fig. [S9](#MOESM1){ref-type="media"}).Figure 4Experimental NR data (markers) and best fits (lines) after incubation for 8 hours with HDL (left) or LDL (right) on h- (upper row) or d- (lower row) SLBs made of DMPC/DMPS (90:10 mol%). The bulk contrasts are dTris buffer (blue), hTris buffer (yellow), and a solvent mixture with the same SLD as silicon (CMSi, green). For clarity, data for CMSi and dTris contrasts are offset by 10 and 100, respectively. Table 2Effects of exposure of supported DMPC/DMPS (90:10 mol%) bilayers to HDL and LDL on bilayer structure and composition.LipoproteinHDLLDLChange in thickness of tail region+1 Å+2 ÅRemoval of lipids from tail^a^22, 26%7, 13%SLD of tail region after incubation4.91 × 10^−6^ Å^−2^4.60 × 10^−6^ Å^−2^% hydrogenated material exchanged with the bilayer26%31%Coverage of surface with lipoprotein^a^2, 6%2, 3%^a^Values for the two different supported bilayers- hydrogenated and deuterated, respectively.
Parameters for the initial SLB and surface (Table [1](#Tab1){ref-type="table"}) were used to fit the data after lipoprotein incubation, keeping the underlying surface fixed whilst allowing the bilayer parameters to vary within limits. In order to restrict the number of free parameters, both h- and d-SLBs were fitted simultaneously for each lipoprotein class. Multiple models were considered, including a simple adsorbed lipoprotein layer on the SLB surface, lipid exchange (change in the SLD of the tail region) and lipid removal (change in hydration of the bilayer). The best fit to the data was found when combining all models. The percentage of lipid removal is seen as a shift in the SLD of the tail region towards the water contrast where the lipids are replaced by water instead of hydrogenous material. Similar models of lipid removal from bilayers have previously been observed upon interaction with biomolecules and nanoparticles using NR, ellipsometry and AFM^[@CR45]--[@CR47]^. The exposure of aliphatic chains to the aqueous solution is energetically unfavourable, however, for multicomponent lipid systems, such as the presently investigated lipoproteins, the energy required to form voids or bilayer patches due to lipid removal could be counterbalanced by preferential transfer of shorter lipid and conic-like fat molecules from the lipoproteins into the SLB. Furthermore, lipid molecules around the fringe of the defect are likely to reorient and/or adjust their packing in order to minimize such aliphatic chain exposure to the aqueous environment. NR does not have the resolution parallel to the interface to detect lipid rearrangement around a small number of pores in the bilayer, instead it is detected as a roughening of the head and tail regions. The fits were further refined by separately fitting the h- and d-SLBs within reasonable error limits (up to 10% difference) to produce a global model for the interaction of the lipoproteins with lipid bilayers. This was necessary as the amount of lipoprotein adsorbed and the percentage of lipids removed differed slightly between parallel experiments. As expected, these two effects were coupled: a higher lipoprotein adsorption to the d-SLB corresponded to a greater removal of lipids from this bilayer. Importantly, the SLD of the hydrogenated tail regions remained constant in the hTris contrast throughout both experiments, ruling out the deposition of proteinaceous material from the lipoprotein. Proteins have higher SLDs than the other components of lipoproteins (lipids, cholesterol or triglycerides) due to the lower hydrogen to carbon ratio and the greater presence of other atoms with higher scattering length such as oxygen, nitrogen and phosphorous. A summary of the fitted model after incubation is given in the Supplementary Information (Fig. [S10](#MOESM1){ref-type="media"}).
Only a small fraction of the lipoproteins bound to the SLB remained after washing with buffer (6% and 2% HDL coverage for d-and h-SLBs, respectively). Thus, quantitatively, the change in SLB composition is very large compared to the amount of HDL adsorbed at the bilayer with 52% of the deuterated lipids removed from the bilayer over 8 hours. This demonstrates that HDL particles effectively act as nanoscopic 'reservoirs' or 'sponges'; each adsorbed lipoprotein particle having a large effect on the bilayer. The SLDs of the lipoproteins were calculated and fixed as 2.02 × 10^−6^ Å^−2^ for HDL and 2.12 × 10^−6^ Å^−2^ for LDL. A change in the SLD of the lipoprotein layer due to increased content of d-lipids only translates to a change in the lipoprotein layer coverage and leaves the SLB fitting parameters largely unaffected.
The fraction of hydrogenated material exchanged with the lipoproteins was calculated from the fitted SLD of the tail region of the d-SLB after incubation with HDL and found to be 26%. It is not possible to distinguish the identity of the molecules inserted into the bilayer using NR. However, due to the low exchange rates of cholesterol ester and protein and the low solubility of triglycerides in phospholipid bilayers, exchange of these lipoprotein components is expected to be minor^[@CR10],\ [@CR48],\ [@CR49]^. It was, however, unambiguous that hydrogenated material lies within the tail region of the bilayer due to contrast variation measurements. From Fig. [1](#Fig1){ref-type="fig"}, one would expect, for 26% exchange, a distinct minimum at 0.05 Å^−1^. This effect, though, is masked by the extra reflection from the adsorbed HDL that overlaps with this feature, smearing out the dip in the dTris contrast. The fraction of hydrogenated material exchanged with the SLB was higher for LDL compared to HDL (31% to 26%, respectively). For both lipoproteins, the number of lipids in the adsorbed lipoprotein layer after washing was not sufficient to account for the hydrogenated material inserted into the bilayer (assuming that the measured layer thickness reflects the actual total thickness of the adsorbed particles on the surface). Probably, a dynamic adsorption process of lipoprotein particles at the lipid bilayer surface takes place since there is a bilayer particle lipid ratio of 2500:1 for HDL and 5500:1 for LDL when taking into account all particles in the surrounding solution.
There was a small difference detected between the h- and d-SLBs in terms of the extent of lipid removal. However, the removal of lipids caused by LDL (7% and 13% for h- and d-SLBs, respectively) was significantly lower than the values calculated for the HDL incubated bilayer (22% and 26% for h- and d-SLBs, respectively).
When modelling the reflectivity data for the d-SLB, a small fringe at 0.1 Å^−1^ could not be accurately reproduced with simple adsorption and exchange. This fringe was clearly seen upon plotting in RQ^[@CR4]^ vs Q (Fig. [4b](#Fig4){ref-type="fig"} inset). This fringe is a weak Bragg peak that arises from correlation of the thicknesses of adsorbed LDL on the surface. The repeat layers were found to be sensitive to coverage and thickness, but not to SLD due to the low coverage in this layer. The peak was best fitted with 5 repeats of 28 Å with 3% coverage and 30 Å with 2% coverage. A single layer of LDL before the corrugation with 35 Å thickness and 4% coverage improved the fit. Such layering could arise from different alignment of LDL particles on the SLB surface. A schematic representation of the modelled changes in the SLB before and after incubation with HDL and LDL is shown in Fig. [5](#Fig5){ref-type="fig"}.Figure 5Schematic illustration showing lipoprotein-mediated uptake and exchange of lipids between supported lipid bilayers and lipoprotein particles. Drawing not to scale. For clarity rearrangement after lipid removal around pores in the membrane has not been depicted, further discussion can be found in the text.
The observed differences in HDL and LDL interaction with cell membrane mimics agree with physiological observations of the roles of high and low density lipoproteins, where the HDL removes lipids and cholesterol and the LDL deposits into the vascular wall causing cholesterol laden macrophages to develop into foam cells^[@CR50]^. Interestingly, both processes (removal and exchange) are detected with HDL and LDL. Along with the lack of lipid and cholesterol exchange promoters (such as Lecithin--cholesterol acyltransferase (LCAT), ATP-binding cassette transporter (ABCA1) and Scavenger receptor class B member 1 (SR-B1)), this indicates that lipid exchange in part is a dynamic, passive process ongoing between the bilayer and particle (as is also the case between bilayers and vesicles)^[@CR8],\ [@CR51],\ [@CR52]^. In this context, our data suggests that lipoprotein class critically influences the extent of exchange and removal and that direct contact between the lipoproteins and the model membrane is required for exchange to take place.
Previous studies for phospholipid exchange between lipoproteins and isolated liver plasma cell membranes labelled with ^14^C choline have found that HDL shows higher levels of lipid exchange compared to LDL^[@CR21]^. The rates of exchange and transfer were seen to be strongly correlated to the type of phospholipid with the largest difference seen for sphingomyelin exchange where HDL exchanged 50% more than LDL^[@CR21]^. Other studies on prostatic human cells present conflicting results and another *in vitro* study on perfused heart tissue showed no exchange at all^[@CR53],\ [@CR54]^. Comparisons are difficult as some experiments are performed *in vivo* and others *in vitro* under very different experimental conditions. For example, the total lipoprotein particle concentration and available lipid concentration differ between the lipoprotein samples. In our experimental set up, a constant number of particles was used in the solid-liquid flow cell. Due to the differences in size and composition, this implies that there is roughly twice the number of lipid molecules in the LDL sample as compared to HDL. In order to clarify the role of the lipoprotein type, a systematic study focusing on varying the concentration of the lipoprotein particles is therefore needed. With this information the studies can further be expanded to include a wide range of factors thought to be linked to the onset of atherosclerosis including the exact lipoprotein structure and composition (nascent vs. mature HDL), oxidation degree (oxidized LDL has been linked to a higher risk of atherosclerosis development)^[@CR55]^, SLB composition (including sphingomyelin, saturated versus unsaturated lipids and cholesterol in the SLB to study preferential uptake), lipoprotein ratio, and the protective effect of HDL on LDL adsorption. Currently, we are expanding our studies to examine the effect of the SLB composition in terms of its surface charge and the content of cholesterol. Initial analysis of the data suggests clear differences in the kinetics for both lipoprotein adsorption and lipid exchange depending on the lipoprotein class and the SLB composition. This work can be further expanded to study the role of lipoprotein specific membrane bound receptors using surface modifications as established recently^[@CR56],\ [@CR57]^.
Conclusions {#Sec7}
===========
Quartz crystal microbalance with dissipation and neutron reflection were used to study adsorption, lipid exchange, and lipid removal kinetics between human lipoproteins and model lipid membranes. Through the use of both hydrogenated and deuterated supported lipid bilayers *in situ*, as well as various isotropic contrasts, it was demonstrated that HDL and LDL interact quite differently with membranes. Thus, HDL, demonstrating anti-atherosclerotic effects in epidemiological studies, was shown to be effective in removing lipids from the bilayer. In contrast, LDL, known as an atherosclerotic risk factor, was shown to deposit more hydrogenous material from the lipoprotein to the SLB. Taken together, the present approach thus allows exchange and other transfer processes in atherosclerosis to be investigated in detail. Importantly, it can be readily extended to interactions with other important components such as the presence of anti-atherosclerotic and other drugs, divalent cations, and oxidative components in atherosclerosis.
Methods {#Sec8}
=======
Materials {#Sec9}
---------
Ultrapure water (18.2 Ω cm^−1^) was used for all experiments. D~2~O was provided by the ISIS neutron source (Rutherford Appleton Laboratory, Didcot, UK). Buffered solutions were made by dissolving a buffer tablet which was pre-adjusted to pH 7.4 in either H~2~O or D~2~O (50 mM Tris with 150 mM NaCl). Chloroform anhydrous ≥99% and calcium chloride dihydrate were purchased from Sigma Aldrich and used as received. Hydrogenated and tail deuterated (1,2-dimyristoyl-*sn*-glycero-3-phosphocholine) DMPC and (1,2-dimyristoyl-*sn*-glycero-3-phospho-L-serine) DMPS were purchased from Avanti Polar Lipids, US. Protein concentration was used as a measure for the concentration of lipoprotein in solution. Bradford Reagent was used for protein determination as previously detailed^[@CR58]^. Phosphate analysis was performed using the method outlined in Rouser *et al*.^[@CR59]^ and showed for the same protein concentration LDL contained twice the number of phospholipid molecules compared to HDL.
Preparation of lipoproteins {#Sec10}
---------------------------
Lipoprotein was prepared as described previously^[@CR60]^. Briefly, plasma from three healthy males was pooled and purified by sequential ultracentrifugation with densities of 1.065 and 1.019 g mL^−1^ for HDL and LDL respectively. The samples were stored in 50% sucrose, 150 mM NaCl, 24 mM EDTA, pH 7.4 at −80 °C. Up to one week before use, the lipoproteins were buffer exchanged to 50 mM Tris, 150 mM NaCl, pH 7.4 (Sephadex G25 PD-10 desalting column) and further purified by size exclusion chromatography (Superose 6 Increase 10/300 GL column) at 25 °C. Each fraction was then stored away from light, at 4 °C, under an inert atmosphere.
Prior to injection the protein concentration was determined by Bradford analysis^[@CR58]^ and the solutions diluted to either 0.132 mg mL^−1^ HDL or 0.1 mg mL^−1^ LDL, these concentrations were chosen to maintain a constant particle concentration in the cell^[@CR61]^.
Preparation of lipid films {#Sec11}
--------------------------
Lipid films consisting of 90 mol% DMPC and 10 mol% DMPS (dissolved in chloroform or a 2:1 mixture with methanol for DMPS) were prepared by drying a lipid mixture of appropriate ratio onto the walls of clean glass vials using a stream of nitrogen at 37 °C for 15 min.
The films were further dried in a vacuum oven overnight and kept at −20 °C until use. For the experiments 2 mL of H~2~O was added to each vial and hydrated in a water bath for at least 1 hour at 40 °C in order to be above the phase transition temperature for both DMPC (h-DMPC 24 °C and d-DMPC 20 °C) and DMPS (h-DMPS 35 °C and d-DMPS 31 °C). Deuterated lipid transition temperatures are lower by 4 °C due to decreased hydrogen bonding^[@CR62]^. All hydrated lipid films were sonicated immediately before injection using a tip sonicator (Hielscher, Germany) intermittently for 5 minutes whilst ensuring the temperature did not rise above 50 °C.
QCM-D {#Sec12}
-----
Experiments were performed using a Q-Sense E4 quartz crystal microbalance. Before the experiment the QCM-D cells and tubings were thoroughly cleaned in a 2% Hellmanex^®^ solution followed by ethanol with sonication and dried in a nitrogen stream. The silicon oxide sensor chip was also cleaned in 2% Hellmanex^®^ and water/ethanol, dried, and plasma cleaned for 1 min before sealing in the cell housing. All experiments were carried out at 37 °C.
The resonance frequencies were acquired in water and once a stable baseline had been reached 1 mL of recently tip sonicated vesicles was mixed with 1 mL 4 mM CaCl~2~ and pumped into the cell at 50 μL min^−1^. After the initial drop in frequency had plateaued the vesicles were washed away in pure water (\~10 min). The solution was then exchanged with h-Tris buffer to form a stable baseline with the lipid bilayer. Lipoprotein samples (0.1 mg mL^−1^ LDL or 0.132 mg mL^−1^ HDL, measured as protein concentration) in Tris buffer were pumped into the cell at 50 μL min^−1^ for 20 minutes. The pump was then stopped and the system allowed to equilibrate for 12 hours. After equilibration the bilayers were washed with tris buffer for 30 minutes at 50 μL min^−1^. Experiments were repeated on consecutive days.
Neutron Reflection {#Sec13}
------------------
NR measurements were performed using home made flow cells. Silicon (111) blocks (80 × 50 × 15 mm) were first treated with dilute piranha solution (5:4:1 H~2~O:H~2~SO~4~:H~2~O~2~) at 80 °C for 15 minutes before UV ozone treatment for a further 10 minutes. The PEEK troughs and O-rings were cleaned with a 2% Hellmanex solution then ultra-pure water with sonication before use. The sample cells were connected to an HPLC pump to allow contrast changes *in situ* and the temperature was kept at 37 °C by a circulating water bath.
Experiments were carried out on the INTER reflectometer at the ISIS neutron source (Rutherford Appleton Laboratory, Didcot, UK)^[@CR63]^. Two incident angles, 0.8° and 3.2°, were used to cover the Q-range of interest (0.01 to 0.25 Å^−1^). Collimating slits were set to give a constant footprint of the neutron beam on the crystal surface of 60 × 35 mm with an experimental resolution (δλ/λ) of 6%.
The silicon oxide surface was first measured in two contrasts, H~2~O and D~2~O, to assess the surface roughness and cleanliness of the crystal. 2 mL of recently tip sonicated vesicles were then mixed with 2 mL 4 mM CaCl~2~ and introduced to the cell by syringe and allowed to incubated for 15 minutes. The excess vesicles were removed by flushing through 5 mL of 2 mM CaCl~2~ solution and 20 mL of buffer. The resulting bilayer was characterised in three solvent contrasts, H~2~O, D~2~O and a mixture that results in a SLD that matches silicon (2.07 × 10^−6^ Å^−2^).
Lipoprotein solutions were introduced to the sample cell at 1 mL min^−1^ by syringe pump via an injection port. The solutions were allowed to equilibrate for 8 hours whilst measuring every 10 minutes for the first hour, then every 2 hours before washing the bilayer. The sample was then characterised again in the same three contrasts as before.
Data sets were fitted using the RasCAL reflectivity fitting software, which calculates multilayer reflectivity using the Parratt formulation^[@CR64]^. Data for each sample (bare surface and deposited bilayer) were fitted simultaneously.
Data Availability {#Sec14}
-----------------
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
Electronic supplementary material
=================================
{#Sec15}
Supplementary Information
**Electronic supplementary material**
**Supplementary information** accompanies this paper at doi:10.1038/s41598-017-07505-0
**Publisher\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors wish to thank ISIS neutron source (RB1610149) for the allocation of beam time and the beamline scientists and support staff, especially Dr. Maximillian Skoda, for support during experiments. We also wish to thank Ann-Margreth Carlsson and Ingrid Söderberg for isolation and Irena Ljungcrantz for purification of the lipoproteins and Habte Ghebrai for analysis of the prepared lipoproteins.
K.L.B. planned and performed QCM-D and NR experiments, analysed NR data, wrote the manuscript. T.K.L. participated in the planning and in the measurement of QCM-D and NR data, performed S.E.C. purification, N.M.R. experiments, assisted in NR analysis, commented and approved the manuscript. S.M. participated in the planning and in the measurement of NR data, performed S.E.C. purification, commented and approved the manuscript. S.M.-H. assisted the NR experiments. G.N.F., E.B. and M.M. participated in the planning of the experiments, commented and approved the manuscript. M.C. acquired the funding, participated in the planning, performance and analysis of the QCM-D and NR experiments, commented and approved the manuscript.
Competing Interests {#FPar1}
===================
The authors declare that they have no competing interests.
|
Q:
How to use the .d.ts typings locally for vscode intellisense in a node.js project?
I am setting up a node.js project that uses a native add-on. The native add-on includes a large number of exported functions. I've setup a typings file (.d.ts) that includes all the function definitions and data etc. that are exported from the native add-on. When I pack all of this up with npm, and install it into the client project, the vscode intellisense picks up all the types and all is well.
When I try to use the typings for a test.js in the same project as the native add-on, the typings are not being picked up, specifically the exported variables; I suspect it has something to do with the way they are exported in the .d.ts, or the naming of the module in the .d.ts.
In the .d.ts, I have the exports listed as;
interface MyI {
Initiate() : void;
}
module 'modulename' {
export var i : MyI;
}
I require the module in the client as (.js file);
var i = require("modulename");
In the test code, I require it as (since I stub it through a index.js file);
var i = require("./index.js");
The index.js in turn looks like;
var i = require("./lib/nativeaddon");
module.exports.i = i;
How to I get vscode to use the typings, locally, for the intellisense when I use the add-on (via the index.js) for the test.js?
A:
To create the typings for vscode intellisense to work for both the "local" case (functional with test.js) and the "global" case (as a node_module), naming the file after the main/entry .js does the trick. In this case the "main" file is index.js, so the typings become index.d.ts.
This does seem natural, but I haven't been able to find the documentation for the vscode intellisense that specifies this as such.
I had previously named the typings after the package/node_module name, packagename.d.js) and kept the "main" (from package.json) as index.js. The "typings" value in the package.json should also match the .d.ts file name.
I suppose a neat alternative to the "index.js" or "main.js", would be to name the main entry point, and the corresponding typings, after the package name.
|
Q:
How to use Search Bar for enter text and filter for selection?
I use textfield for enter field of my web service.For example I entered StationID textfield=35016.
I have total 770 StationID and every station have a name. stationId 35016 name is New Istanbul LTD STI.I want when I entered New to search bar it must will list to me New Istanbul LTD.STI.Than I will select for sending web service call.
How can I do for search and select searching field.This code for Textfield.How to change for search bar ? Thank you.
in .h file
#import <UIKit/UIKit.h>
@interface AMDViewController : UIViewController<UITextFieldDelegate,NSXMLParserDelegate>
@property (unsafe_unretained, nonatomic) IBOutlet UITextField *StationID;
@end
in .m file
enter code here
#import "AMDViewController.h"
@interface AMDViewController ()
{
NSMutableData *webData;
NSXMLParser *xmlParser;
NSMutableString *retornoSOAP;
BOOL teveRetorno;
@end
@implementation AMDViewController
@synthesize StationID;
}
-(IBAction)calcularTemperatura:(UIButton *)sender{
NSString *mensagemSOAP= [NSString stringWithFormat:@"<?xml version=\"1.0\" encoding=\"utf-8\"?>\n"
"<soap:Envelope xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" xmlns:xsd=\"http://www.w3.org/2001/XMLSchema\" xmlns:soap=\"http://schemas.xmlsoap.org/soap/envelope/\">\n"
"<soap:Body>\n"
"<Details xmlns=\"http://tempuri.org/\">\n"
"<StationID>%@</StationID>\n"
"<StationName>%@</StationName>\n" //StationName is here in web sevrice
"</Details>\n"
"</soap:Body>\n"
"</soap:Envelope>\n",StaionID.text];
NSLog(@"SOAP msg = \n%@\n\n", mensagemSOAP);
NSURL *url = [NSURL URLWithString:@"http://webservice/sample.asmx"];
NSMutableURLRequest *theRequest = [NSMutableURLRequest requestWithURL:url]; NSString *tamanhoMensagem = [NSString stringWithFormat:@"%d", [mensagemSOAP length]];
[theRequest addValue:@"text/xml; charset=utf-8" forHTTPHeaderField:@"Content-Type"];
[theRequest addValue: @"http://tempuri.org/Details" forHTTPHeaderField:@"SOAPAction"];
[theRequest addValue:tamanhoMensagem forHTTPHeaderField:@"Content-Length"];
[theRequest setHTTPMethod:@"POST"];
[theRequest setHTTPBody:[mensagemSOAP dataUsingEncoding:NSUTF8StringEncoding]];
NSURLConnection *conexao = [[NSURLConnection alloc] initWithRequest:theRequest delegate:self];
if(conexao){
webData = [NSMutableData data];
}else{
NSLog(@"Connection Error.");
}
}
-(void)parser:(NSXMLParser *)parser didStartElement:(NSString *)elementName namespaceURI:(NSString *)namespaceURI qualifiedName:(NSString *)qName attributes:(NSDictionary *)attributeDict{
if ( [elementName isEqualToString:@"StationID"] ) {
if (!retornoSOAP) {
retornoSOAP = [[NSMutableString alloc] init];
}
teveRetorno = YES;
}
}
- (void)parser:(NSXMLParser *)parser didEndElement:(NSString *)elementName namespaceURI: (NSString *)namespaceURI qualifiedName:(NSString *)qName{
if ( [elementName isEqualToString:@"StationID"] ) {
StaionTotalSalesTodayLabel.text = retornoSOAP;
retornoSOAP = nil;
teveRetorno = NO;
}
A:
You have to use uisearchbar control and its delegate methods.
Here is the link that may help you
http://www.youtube.com/watch?v=P2yaZXn4MU0
|
Q:
Copy ListBox to Clipboard, return carrier not conserved
In a C++ application using MFC, i would like to be able to copy the entire CListBox content to clipboard.
I found a function which copy the content, neverheless, the return carrier aren't conserved.
I look with an HexEditor, and it appear there is $0A instead of $0D&$0A.
Here is my code :
CListBox * myListBox = (CListBox *)GetDlgItem(IDC_LIST_RESULT);
CString sContents = _T("");
CString temp = _T("");
int NumberOfSelections = 0;
NumberOfSelections = myListBox->GetCount();
for(int Selection = 0; Selection <= NumberOfSelections-1; Selection++)
{
myListBox->GetText(Selection, temp);
sContents += temp;
sContents +="\n";
}
if (OpenClipboard())
{
HGLOBAL clipbuffer;
char * buffer;
if (EmptyClipboard())
{
clipbuffer = GlobalAlloc(GMEM_DDESHARE, sContents.GetLength() + 1);
buffer = (char*)GlobalLock(clipbuffer);
CStringA ansiString(sContents);
size_t cbString = strlen(ansiString) + 1;
strcpy_s(buffer, cbString, ansiString);
GlobalUnlock(clipbuffer);
if (SetClipboardData(CF_TEXT, clipbuffer) == NULL)
{
CString msg;
msg.Format(_T("Unable to set Clipboard data, error: %d"), GetLastError());
AfxMessageBox(msg);
}
else
AfxMessageBox(_T("Successfully copied selected laps to clipboard"));
}
else
AfxMessageBox(_T("Unable to empty Clipboard"));
CloseClipboard();
}
else
AfxMessageBox(_T("Unable to open Clipboard"));
// TODO: ajoutez ici le code de votre gestionnaire de notification de contrôle
I use unicode configuration with Visual Studio 2013.
Anyone have some ideas ?
Thanks a lot,
Best regards,
Nixeus
A:
There only a \n because that's what you put in the clipboard.
sContents +="\n";
It should be
sContents +="\r\n";
|
OLYMPUS MONS, Mars - Driving a car on Mars is not only a technical challenge, it's also an expensive hobby. This is what NASA had to learn the hard way these days. The Mars rover "Spirit" is facing a period of "hibernation" for budget reasons, while there are no technical obstacles to continue the mission.
"We have built an extremely intelligent, self-controlling robot for this mission," a NASA spokesman said, "but we misjudged its inclination to spend recklessly." The self-driving robot regularly hits the bars and casinos on the red planet, checks into the fanciest hotels, and also has a soft spot for precious stones. And then, gas prices on Mars presently are going through the roof as well.
NASA bankers are about to pull the plug of the smart machine. Before sending any more intelligent devices to other planets, they demand development of a "cheapskate" module for them.
The Spirit rover is also wanted for questioning in the mysterious death of its twin, Opportunity, who was found at the bottom of Valles Marineris three weeks ago. Police officials say this was a murder that was made to look like a suicide.
"No way that rover offed himself," said Chief Investigator Marvin Martin. "There were clear signs that it was hit from behind as it was looking over the edge. There are clear scrape marks that weren't caused by a plunge off a cliff. This makes me very angry."
Spirit has denied any involvement in the death, although he recently got a new paint job, potentially destroying all evidence of his culpability. |
INTRODUCTION {#s0}
============
Fungal infections affect in excess of a billion people, resulting in approximately 11.5 million life-threatening infections and more than 1.5 million deaths annually ([@B1]). Candida auris is an emerging fungal pathogen that has attracted considerable attention because of its ability to cause infections that are difficult both to diagnose and to treat ([@B2]). It has been responsible for a number of nosocomial outbreaks worldwide through its ability to persistently colonize and be transmitted between patients and the environment ([@B3][@B4][@B6]). Despite the unprecedented global emergence of this organism, relatively little is known about the molecular basis of its pathogenicity and antifungal resistance phenotype. The resistance profile is well documented, with \>90% of isolates intrinsically resistant to fluconazole. Resistance to other azoles, polyenes, and echinocandins has also been reported ([@B4]). Alarmingly, 41% of isolates have been shown to be multidrug resistant, with 4% demonstrating pan-drug resistance ([@B4]). Hot spot mutations in *ERG11* and *FKS1* have been identified as resistance mechanisms in azole- and echinocandin-resistant strains, respectively ([@B7], [@B8]).
*Candida* biofilms represent an important clinical entity associated with adaptive resistance to many antifungals and are linked to excess morbidity and mortality ([@B9][@B10][@B11]). Although Candida albicans is regarded as the primary biofilm-forming pathogen within the genus, there are increasing interest in and evidence for non-Candida albicans-species biofilms ([@B12], [@B13]), particularly those of C. auris. Clinically, C. auris has been isolated from a number of sites, including wounds, line tips, and catheters, suggestive of the organism existing within a biofilm lifestyle in the host ([@B14], [@B15]). We recently described the ability of C. auris to form antifungal-resistant biofilms, against all 3 main classes of antifungals ([@B16]), and yet the mechanisms underlying this phenotype remain unknown. The speed of discovery in this emerging pathogen has certainly been hindered by the lack of robust sequence information. Initial sequencing efforts provided a draft C. auris genome; however, these reads were poorly aligned to other *Candida* spp. and inconsistently annotated ([@B17]). More recently, complete and functionally annotated genome assemblies have been created, allowing the analysis of the functional capacity of the genome to be studied under clinically relevant conditions ([@B18]). Biofilm-associated resistance is a complex and multifaceted phenomenon that has been described in a number of fungal pathogens. Various resistance mechanisms exist, predominately associated with the extracellular matrix (ECM), overexpression of drug targets, and efflux pumps ([@B19]). Given the lack of understanding of biofilm formation and resistance mechanisms in C. auris, we therefore aimed to investigate these mechanisms using a transcriptomics approach.
RESULTS {#s1}
=======
*Candida auris* biofilms exhibit temporal antifungal resistance. {#s1.1}
----------------------------------------------------------------
Mature Candida auris biofilms have been shown to be resistant to antifungals that are readily active against their planktonic equivalents ([@B16]). We therefore investigated the temporal effect of biofilm formation on the susceptibility to all three major classes of antifungals. As demonstrated in [Fig. 1A](#fig1){ref-type="fig"}, the maturation of C. auris biofilms was shown to correlate with decreased susceptibility to each antifungal agent. When assessed planktonically, the median MIC for the four isolates of miconazole was 1 µg/ml, that of micafungin was \<0.25 µg/ml, and that of amphotericin B was 0.5 µg/ml (range, 0.125 to 0.5 µg/ml). After 4 h of biofilm development, no increases in resistance were observed against micafungin (MIC, \<0.25 µg/ml); however, the median MIC increased 16-fold to 16 µg/ml (range, 16 to 32 µg/ml) for miconazole and 4-fold to 2 µg/ml for amphotericin B (range, 1 to 4 µg/ml). As the biofilm matured to 12 h of growth, 2-fold increases in median MIC were shown for miconazole (range, 16 to 64 µg/ml) and amphotericin B (range, 2 to 4 µg/ml). Interestingly, the MIC was shown to significantly increase for micafungin (range, 1 to \>128 µg/ml) after 12 h. After 24 h, no further increase in MIC was observed for amphotericin B. However, both miconazole and micafungin MICs were increased 2-fold to 64 µg/ml and \>128 µg/ml, respectively.
{#fig1}
*Candida auris* transcriptome assembly. {#s1.2}
---------------------------------------
Given the temporal patterns of biofilm-associated resistance, we undertook a transcriptional profiling approach to understand the mechanisms governing antifungal biofilm resistance ([Fig. 2](#fig2){ref-type="fig"}). Sequencing of samples using Illumina HiSeq produced around 414 million single-end reads of 50-bp length. Following processing, the number of reads was reduced by 3 million through trimming and quality control stages. All sequenced sample reads were then assembled into an \~11.5-Mb transcriptome which consisted of 5,889 identified Trinity transcripts and 5,848 genes based on the longest isoform of transcripts. At least half of the assembled sequenced bases were found on contigs of a length of 3,488 bp (*N*~50~) ([Table 1](#tab1){ref-type="table"}). The completeness and quality of the C. auris transcriptome were assessed with BUSCO (Benchmarking Universal Single-Copy Orthologs) against Ascomycota (94%), Saccharomyceta (91.4%), and Saccharomycetales (91.7%) gene sets. Very small percentages of duplicate, fragmented, and missing genes were observed in each of the gene sets ([Table 2](#tab2){ref-type="table"}).
{#fig2}
######
Summarized statistics for transcriptome assembly of Candida auris, alignment rate of raw reads to transcriptome, and summary of Trinotate functional annotation
Category Value
------------------------------------------- ---------------------
No. of reads
Total 414,364,539
After trimming 411,626,529
Total no. of assembled bases 11,593,681
GC content, % 45.35
Total no. by Trinity
"Genes" 5,848
"Transcripts" 5,889
Contig (bp)
*N*~50~ 3,488
Median 1,308
Avg 1,983
No. of reads aligned (%)
1 time 393,124,946 (95.51)
\>1 time 9,368,727 (2.28)
Overall 402,493,673 (97.78)
Functional annotation, no. of transcripts
Swiss-Prot matches, BLASTx 3,200
Swiss-Prot unique proteins, BLASTx 3,176
Swiss-Prot matches, BLASTp 3,041
Swiss-Prot unique proteins, BLASTp 3,019
TMHMM 701
SignalP 202
Gene Ontology 3,085
KEGG 2,889
######
Assessment of Candida auris transcriptome assembly by Benchmarking Universal Single-Copy Orthologs (BUSCO)
\% genes Ascomycota Saccharomyceta Saccharomycetales
---------------------- ------------ ---------------- -------------------
Complete 94 91.4 91.7
Complete single copy 93.4 90.5 90.9
Complete duplicated 0.6 0.9 0.8
Fragmented 3.4 4.8 4.6
Missing 2.6 3.8 3.7
Total no. of genes 1,315 1,759 1,711
Identification by sequence homology searches with BLASTx function yielded annotation of 54% of Trinity transcripts and 54% of unique "genes." Identification of protein sequences with BLASTp, against TransDecoder-identified open reading frames (ORFs) and potential coding sequences, gave functional annotation matches with 51% of the transcripts and 41% of unique "genes" ([Table 1](#tab1){ref-type="table"}). The presence of known signal peptides, functional protein domains, and protein topology was discerned by searches against the SignalP and TMHMM databases, respectively. Of the predicted proteins, 202 sequences were predicted to have signal peptides and 701 transmembrane protein topologies were predicted.
Additional annotation was performed via the software BLAST2GO, which obtains BLAST hits that are used to retrieve and map gene ontology (GO) and KEGG terms. It also utilizes InterProScan, which acquires functional annotation of protein sequences from EBI's InterPro databases (<https://www.ebi.ac.uk/interpro/>). These databases are a consortium of online databases that include PANTHER, Pfam, and SUPERFAMILY ([@B20]). Both the Trinotate and BLAST2GO annotation files are supplied as [Data Set S1](#dataS1){ref-type="supplementary-material"} in the supplemental material.
10.1128/mSphere.00334-18.5
Trinotate and BLAST2GO annotation. Download DATA SET S1, XLSX file, 12 MB.
Copyright © 2018 Kean et al.
2018
Kean et al.
This content is distributed under the terms of the
Creative Commons Attribution 4.0 International license
.
BLAST2GO searches were performed with a fungus taxonomical filter, which annotated 1,157 genes with BLAST and an additional 4,365 genes from the InterPro databases. InterPro and BLAST-derived GO terms were merged to give a total of 9,504 GO annotations assigned to 2,479 genes. These annotations were distributed among three main GO categories, biological process (3,633, 38%), cellular component (3,116, 33%), and molecular function (2,755, 29%) ([Fig. S1](#figS1){ref-type="supplementary-material"}). InterProScan was able to classify Trinity transcripts according to superfamilies based on known structures. The best-represented superfamilies were the P-loop-containing nucleoside triphosphate hydrolase (236 genes), the major facilitator superfamily (MFS) (113 genes), Armadillo-type fold (102 genes), and protein kinase-like superfamily (90 genes) ([Fig. S2](#figS2){ref-type="supplementary-material"}). From annotation against the available databases, there were 6 major enzyme classes represented, which included hydrolyases (290 genes), transferases (150 genes), oxidoreductases (88 genes), ligases (21 genes), lyases (22 genes), and isomerases (15 genes) ([Fig. S3](#figS3){ref-type="supplementary-material"}).
10.1128/mSphere.00334-18.1
Gene Ontology term distribution. GO-based annotation was performed by BLAST2GO and InterProScan, which functionally classified transcripts into terms associated with biological process (BP), cellular component (CC), and molecular function (MF). Download FIG S1, TIF file, 2.1 MB.
Copyright © 2018 Kean et al.
2018
Kean et al.
This content is distributed under the terms of the
Creative Commons Attribution 4.0 International license
.
10.1128/mSphere.00334-18.2
Superfamily distribution of assembled transcripts. Assembled transcripts were submitted to an InterProScan search against the SUPERFAMILY distribution protein database for functional classification, with 479 annotations against identified sequences. Download FIG S2, TIF file, 2.1 MB.
Copyright © 2018 Kean et al.
2018
Kean et al.
This content is distributed under the terms of the
Creative Commons Attribution 4.0 International license
.
10.1128/mSphere.00334-18.3
Functional characterization of potential enzyme transcripts in Candida auris. Assembled transcripts were identified by BLAST2GO via enzyme code mapping with transcripts distributed across 6 major enzyme classes. Download FIG S3, TIF file, 2.1 MB.
Copyright © 2018 Kean et al.
2018
Kean et al.
This content is distributed under the terms of the
Creative Commons Attribution 4.0 International license
.
DE and functional annotation of C. auris biofilms. {#s1.3}
--------------------------------------------------
Differential expression (DE) analysis was performed to investigate the transcriptional changes observed with biofilm development. Multivariate analysis by principal-component analysis (PCA) demonstrates variance between the different time points; 0 h shows the greatest variance from the other biofilm time points. In addition, there is also some variance between biofilms at 4, 12, and 24 h ([Fig. 3A](#fig3){ref-type="fig"}). DE analysis demonstrated that 791 and 464 genes were upregulated in biofilm formation and planktonic cells, respectively, with a minimum 2-fold change ([Fig. 3A](#fig3){ref-type="fig"}). Phase-dependent differential expression of these upregulated genes is illustrated in the Venn diagram in [Fig. 3B](#fig3){ref-type="fig"}, with the downregulated genes shown in [Fig. 3C](#fig3){ref-type="fig"}; individual genes are described in [Data Set S2](#dataS2){ref-type="supplementary-material"}. Of these biofilm-upregulated genes, selected genes involved in antifungal resistance and biofilm-associated mechanisms are listed in [Table 3](#tab3){ref-type="table"}. Glycosylphosphatidylinositol (GPI)-anchored cell wall genes, including *IFF4*, *CSA1*, *PGA26*, and *PGA52*, were upregulated at all time points of biofilm formation, highlighting their potential role within cellular adhesion ([Table 3](#tab3){ref-type="table"}). Two further adhesins, *HYR3* and *ALS5*, were also shown to be upregulated but only in mature biofilms ([Table 3](#tab3){ref-type="table"}). As the biofilm developed into intermediate and mature stages, a number of genes encoding efflux pumps were upregulated, including *RDC3*, *SNQ2*, *CDR1*, and *YHD3*. In addition, *MDR1* was shown to be upregulated at the 24-h time point ([Table 3](#tab3){ref-type="table"}). To understand the functional processes related to differentially expressed genes, a cutoff of 2-fold upregulation (adjusted *P* value of \<0.05) was used for gene ontology (GO) analysis comparing planktonic cells to 24-h biofilms. The 278 differentially expressed genes were assigned to 28 GO terms with an overenrichment *P* value of \<0.05, comprising 13 biological processes, 9 cellular components, and 6 molecular functions, and contained a number of differentially expressed functional categories ([Fig. 4A](#fig4){ref-type="fig"}). Included within these GO terms were transmembrane transport, within which several ATP-binding cassette (ABC) and major facilitator superfamily (MFS) transporters were highly upregulated in C. auris biofilms ([Fig. 4B](#fig4){ref-type="fig"}).
10.1128/mSphere.00334-18.6
Differentially expressed genes. Download DATA SET S2, XLSX file, 0.3 MB.
Copyright © 2018 Kean et al.
2018
Kean et al.
This content is distributed under the terms of the
Creative Commons Attribution 4.0 International license
.
{#fig3}
######
Upregulated biofilm- and resistance-associated genes
-----------------------------------------------------------------------------
Gene\ Function Fold change compared to\
identifier planktonic cells (log~2~)
------------ ---------------------- --------------------------- ------ ------
*IFF4* Adhesion 2.29 5.01 3.62
*PGA26* Adhesion 2.02 3.90 2.55
*PGA52* Adhesion 2.22 2.38 2.42
*CSA1* Adhesion 3.87 6.47 6.43
*PGA7* Adhesion 3.94 4.82
*HYR3* Adhesion 2.06
*ALS5* Adhesion 3.82
*RDC3* Efflux pump 4.29 3.91
*SNQ2* Efflux pump 2.63 3.42
*CDR1* Efflux pump 2.30 3.19
*YHD3* Efflux pump 2.14 2.15
*MDR1* Efflux pump 2.3
*KRE6* Extracellular matrix 3.92 3.09
*EXG* Extracellular matrix 2.69 2.26
*SAP5* Hydrolytic enzyme 2.19
*PLB3* Hydrolytic enzyme 2.13
-----------------------------------------------------------------------------
{#fig4}
Efflux pumps play a primary role in antifungal resistance in C. auris biofilms. {#s1.4}
-------------------------------------------------------------------------------
Transcriptional analysis and function annotation revealed a significant upregulation of a number of drug efflux pumps, from both ABC and MFS transporters. To confirm the role of these membrane proteins within biofilms, we assessed efflux pump activity. Both 12- and 24-h biofilms exhibited increased efflux compared to planktonic cells, with 4-h biofilms below the detectable limit of the assay. Efflux from 12-h biofilms was 2.21-fold (*P* \< 0.05) greater than that from planktonic cells, with a 2.38-fold increase shown in 24-h biofilms (*P* \< 0.005). No statistical differences were observed between 12- and 24 h-biofilms ([Fig. 5](#fig5){ref-type="fig"}). Interestingly, efflux pump activity is shown to be constitutively expressed within biofilms, with no induction observed in response to azole antifungals ([Fig. S4](#figS4){ref-type="supplementary-material"}).
10.1128/mSphere.00334-18.4
Azole exposure does not induce Candida auris efflux pump activity. Following biofilm growth for 24 h, C. auris biofilms were treated with a range of fluconazole concentrations (0 to 256 µg/ml) for 4 h before efflux pump activity was quantified using the Ala-Nap assay. Download FIG S4, TIF file, 2.1 MB.
Copyright © 2018 Kean et al.
2018
Kean et al.
This content is distributed under the terms of the
Creative Commons Attribution 4.0 International license
.
{#fig5}
Given the increased activity of efflux pumps in biofilms, we then assessed the contribution of these transporters to fluconazole sensitivity ([Table 4](#tab4){ref-type="table"}). When biofilms were incubated for 12 h in the presence of fluconazole, the sessile MIC~50~ (SMIC~50~) ranged between 32 and \>128 µg/ml. However, when also grown in the presence of fluconazole and an efflux pump inhibitor (EPI), the SMIC~50~ ranged between 2 and 16 µg/ml for all isolates, ranging from a 4- to 16-fold increase in susceptibility. The same trend was observed for 24-h biofilms, with the SMIC~50~ range between 64 and \>128 µg/ml for fluconazole-only treatment, with 2- to 8-fold reductions observed with coincubation with the EPI (SMIC~50~, 8 to 64 µg/ml).
######
Inhibition of efflux pumps increases azole susceptibility
Isolate no. Fluconazole SMIC~50~ (μg/ml) at time:
------------- --------------------------------------- ------- ---- ---- ------- ----
NCPF8971 16 64 4 16 \>128 ≥8
NCPF8973 2 32 16 8 64 8
NCPF8984 16 \>128 ≥8 64 \>128 ≥2
NCPF8990 8 32 4 16 64 4
EPI, efflux pump inhibitor.
DISCUSSION {#s2}
==========
The rapid and simultaneous emergence of the pathogenic yeast C. auris, combined with its reported recalcitrance to all three major classes of antifungals, has led to a concerted response by the medical mycology community to understand and define the mechanisms underpinning its pathogenicity and resistance. Although preliminary investigations have investigated genetic point mutations promoting resistance ([@B7], [@B8]), as well as a number of efflux pumps identified within its genome ([@B17], [@B18]), there are still substantial gaps remaining in our understanding. Moreover, irrespective of these defined chromosomally derived resistance characteristics, adaptive resistance mechanisms associated with environmental stressors are likely to be a key contributor to its success as a pathogen in both the host and the environment. We have recently reported how C. auris exhibits enhanced pathogenicity and resistance, both *in vitro* and *in vivo*, and that the biofilm phenotype is instrumental in its lifestyle ([@B14], [@B16], [@B21], [@B22]), and moreover, in its ability to survive and persist in the nosocomial environment, increasing the probability of causing outbreaks. We have recently reported that adherent C. auris cells display substrate-dependent susceptibility to clinically relevant concentrations of hospital disinfectants ([@B22]) and that these biofilms were shown to be resistant to chlorhexidine and hydrogen peroxide, displaying a less susceptible phenotype than C. albicans and *Candida glabrata* ([@B21]). Here, we report for the first time that efflux-based resistance mechanisms play an important role in biofilm-mediated resistance in C. auris and that conserved biofilm-related genes are temporally observed, as illustrated in [Fig. 6](#fig6){ref-type="fig"}.
{#fig6}
To investigate this, we undertook an RNA sequencing-based approach for the analysis of C. auris biofilm development, as well as profiling genes associated with resistance and virulence mechanisms. Assembly of the transcriptome using Trinity software has allowed us to construct a specific reference for our samples of interest. Additionally, annotation via numerous methods has allowed for an in-depth functional characterization of the organism. Annotation of homologs, predicted protein domains, and gene ontological classifications further enhances our ability to interpret mechanisms that differentiate C. auris under different conditions. This annotated transcriptome has been highly instrumental in expression analysis and elucidation of virulence mechanisms of C. auris in this and forthcoming studies.
The initiation of biofilm formation depends on an initial adherence phase of colonization of a specific surface before subsequent proliferation to promote disease. A number of GPI-linked cell wall proteins were upregulated at the early biofilm time point, highlighting their role in the initial adherence stage. In C. albicans, IFF4 and CSA1 have been shown to be involved in adherence to both mucosal and abiotic substrates, as well as cell-cell cohesion ([@B23][@B24][@B25]). Transcriptional studies from Fox et al. identified *IFF4* as a member of a group of 10 adhesion genes that are induced at the later stages of biofilm formation and hypothesized its role in mediating cell-cell contact ([@B26]). Interestingly, an *iff4*Δ null mutant displayed decreased adhesion at an early stage of biofilm formation, as well as attenuated virulence ([@B27]). Both studies collectively highlight its function throughout biofilm formation.
In C. albicans, members of the agglutinin-like sequence (ALS) proteins play a key role in the adherence of the organism, predominantly through *ALS3* ([@B28], [@B29]). A recent study identified that members of this cell wall protein family detected in C. albicans are not found in C. auris ([@B18]). Our analysis revealed that orthologs of only two members, *ALS1* and *ALS5*, were represented within the C. auris transcriptome, with the latter upregulated within mature biofilms. Further examination of cell wall protein families by Muñoz et al. failed to reveal any highly expanded families ([@B18]). It is therefore likely that a less reliant ALS-dependent adherence mechanism exists for C. auris. Moreover, the gene encoding candidapepsin-5, commonly known as *SAP5* in C. albicans, was shown to be upregulated in mature biofilms. This protease is predominantly associated with its role in invasive infection ([@B30]). Indeed, studies have identified its increased expression in biofilm-associated infections ([@B31]), with *sap5*Δ/Δ strains demonstrating a less adherent phenotype, therefore highlighting its potential as a promising biofilm biomarker ([@B32]).
One of the most defining characteristics of biofilms is their recalcitrance to antimicrobial agents. As described in other *Candida* species, biofilm-associated drug resistance comprises a number of different mechanisms that coordinate with one another through the various phases of biofilm development ([@B33]). An underlying mechanism across *Candida* spp. is the upregulation of efflux pumps within biofilm-associated cells ([@B34][@B35][@B36]). Planktonically, C. auris isolates displayed up to 15-fold-higher ABC transporter activity than C. glabrata isolates ([@B15]), highlighting a potential intrinsic azole resistance mechanism. Ramage et al. demonstrated that expression of *CDR1* and *MDR1* was increased within mature C. albicans biofilms compared to their planktonically grown equivalents, and yet deletion of these genes had no effect on the susceptibility of mature biofilms ([@B37]). Indeed, temporal efflux pump analysis revealed that efflux pump mutants were more susceptible to fluconazole treatment than their parental strain at early phases of biofilm development ([@B36]), as also shown in other fungal pathogens, such as Aspergillus fumigatus ([@B38]). Our own temporal analysis of C. auris biofilms revealed that efflux pumps were upregulated at intermediate and mature phases of development, unlike other species, though they did not appear to be inducible following azole exposure. This is in contrast to analysis of C. glabrata biofilms exposed to azole treatment, where upregulation of genes encoding ABC transporters was observed ([@B35]). Muñoz et al. recently analyzed the transcriptional profile of planktonic C. auris in response to azole and polyene antifungals ([@B18]). After exposure of a resistant C. auris strain to amphotericin B, almost 40 genes were shown to be differentially expressed. These included genes involved in iron transport that have previously been described in C. albicans to be involved in its response to amphotericin B ([@B39]). Three of these genes (*SIT1*, *PGA7*, and *RBT5*) were shown to overlap within our own biofilm data set, indicating that these may play an additional role in our observed polyene resistance.
A further key mechanism of *Candida* biofilm resistance is the formation of the ECM, which functions to provide stability and sequestration of drugs from the biofilm, as well as protection from environmental stressors ([@B40]). Recent studies have now identified that various *Candida* spp. conserve a constitutive polysaccharide backbone that functions to impede antifungal delivery, and yet the composition of the ECM varies between species ([@B41], [@B42]). Although its composition remains unknown, it could be hypothesized that C. auris ECM would be similar to that of C. glabrata, given the yeast cell biofilm phenotype. Temporal analysis has shown that the formation of the ECM is time dependent and associated with intermediate and maturation phases of biofilm formation ([@B43]). Our data suggest that this is similar in C. auris, with increased expression of *KRE6* and *EXG*, a glucan-1,3-beta-glucosidase and a close ortholog of *XOG1* in C. albicans, respectively, two genes involved in matrix formation ([@B44], [@B45]).
Given the alarming global emergence of antifungal resistance, the requirement for new antifungals is pivotal ([@B46]). Drug efficacy and development have plateaued in recent years, yet an encouraging number of molecules remain within the antifungal pipeline ([@B47], [@B48]). Several studies have assessed the positive efficacy of novel compounds, including APX001, CD101, SCY078, and ceragenins, against C. auris ([@B49][@B50][@B52]), which may widen the spectrum of active agents against emerging resistant species. These active agents are both expansions of current drug targets, such as 1,3-β-glucan synthase inhibitors (CD101 and SCY078), and novel targets, such as GPI protein inhibitors (APX001). All of these compounds demonstrated significant *in vitro* activity against planktonic forms of C. auris, with APX001 also demonstrating enhanced *in vivo* efficacy compared to anidulafungin ([@B51], [@B53]). Although these preliminary data are very promising, there are limited studies evaluating their effect against sessile C. auris. The 1,3-β-glucan synthase inhibitor SCY078 was shown to significantly reduce biofilm thickness and metabolic activity after a prolonged 48-h exposure ([@B54]). Furthermore, the CSA-44 and CSA-131 ceragenins, a class of antimicrobial peptides, also demonstrated antibiofilm activity, although the concentrations needed were 4- to 64-fold greater than the planktonically active equivalent ([@B52]). APX001 is a first-in-class compound that acts by blocking GPI synthesis through inhibition of the GPI-anchored cell wall transfer protein 1 (Gwt1). Although no such studies have been performed, perhaps then APX001 is the most attractive antibiofilm target, given our identified function of GPI-anchored proteins in C. auris biofilm formation.
Given that we can now genetically manipulate this pathogenic yeast ([@B55], [@B56]), future work analyzing the functional roles and processes of specific genes and proteins will further enhance our understanding of biofilm-associated pathogenicity and resistance. Unraveling the key factors that regulate the transcriptional network that exists for C. auris, similar to those studies in C. albicans and Candida parapsilosis ([@B26], [@B57]), may provide translational insights into novel avenues for therapeutic targets for biofilm-associated infections. We have shown that efflux pumps are important during biofilm development, and this may explain why this seemingly innocuous yeast is able to survive, persist, and cause continued problems within the hospital setting.
MATERIALS AND METHODS {#s3}
=====================
Microbial growth and standardization. {#s3.1}
-------------------------------------
Four C. auris clinical isolates were used throughout this study (NCPF8971, NCPF8973, NCPF8984, and NCPF8990) ([@B58]). Isolates were stored in Microbank vials at −80°C prior to use, before they were subcultured onto Sabouraud dextrose agar (SAB \[Sigma, Dorset, United Kingdom\]) and incubated at 30°C for 48 h. Isolates were propagated overnight in yeast peptone dextrose (YPD) medium (Sigma, Dorset, United Kingdom), before washing with centrifugation as previously described ([@B59]). Cells were then standardized to 1 × 10^6^ cells/ml in RPMI 1640 medium, and biofilms were grown in microtiter plates, 75-cm^2^ tissue culture flasks, or Thermanox coverslips for 4, 12, and 24 h at 37°C.
Characterization of biofilm formation. {#s3.2}
--------------------------------------
Isolates were standardized as described above and grown for 4, 12, and 24 h at 37°C. Following growth, biofilms were washed with phosphate-buffered saline (PBS; Sigma, Dorset, United Kingdom), and biomass was quantified using the crystal violet assay, as previously described ([@B59]). In addition, biofilm composition was analyzed using propidium monoazide (PMA) quantitative PCR (qPCR), a method able to differentiate live cells from a population ([@B60]). Samples were prepared as previously described ([@B60]), before sonication in 1 ml of PBS at 35 kHz for 10 min in an ultrasonic water bath to remove and disaggregate the biofilm ([@B61]). After sonication, samples were incubated in the dark with 50 µM PMA (Cambridge BioScience, Cambridge, United Kingdom) for 10 min to allow uptake of the dye. All samples were then exposed for 5 min to a 650-W halogen light before DNA was extracted using the QIAamp DNA minikit, per the manufacturer's protocol (Qiagen, Crawley, United Kingdom). One microliter of extracted DNA was then added to a master mix containing Fast SYBR Green master mix, RNase-free water, and 10 µM C. auris-specific forward and reverse primers (forward, CGCACATTGCGCCTTGGGGTA; reverse, GTAGTCCTACCTGATTTGAGGCGAC) ([@B62]). Real-time qPCR was then used to enumerate the total of number of live cells from within the biofilm, using the following thermal profile: 50°C for 2 min and 95°C for 2 min, followed by 40 cycles of 95°C for 3 s and 60°C for 30 s. Colony-forming equivalents (CFE) were then calculated based upon a standard curve of serially extracted DNA ranging from 1 × 10^8^ to 1 × 10^4^ cells/ml.
Biofilm visualization. {#s3.3}
----------------------
Biofilms were standardized and grown on Thermanox coverslips (Fisher Scientific, Loughborough, United Kingdom) as described above. At selected time points, biofilms were washed with PBS before processing for scanning electron microscopy (SEM). Biofilms were fixed in 2% paraformaldehyde, 2% glutaraldehyde, 0.15 M sodium cacodylate, and 0.15% (wt/vol) alcian blue, before being processed as previously described ([@B59]). Biofilms were then sputter coated in gold before being viewed under a JEOL-JSM-6400 microscope.
Planktonic and sessile susceptibility testing. {#s3.4}
----------------------------------------------
Planktonic MICs (pMICs) were determined visually using the Clinical and Laboratory Standards Institute M27-A3 broth microdilution method ([@B63]). Standardized cells were treated with serial 2-fold dilutions of miconazole nitrate (0.25 to 128 mg/liter), micafungin (0.25 to 128 mg/liter), and amphotericin B (0.063 to 32 mg/liter). In addition, biofilms were grown for 4, 12, and 24 h as described above before treatment with the same concentrations as planktonic cells. Sessile MICs (sMICs) were determined using the XTT \[2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide salt\] metabolic reduction assay ([@B64]). The sMIC was calculated as the concentration leading to 80% reduction in XTT colorimetric readings in comparison to an untreated positive control.
RNA extraction and sequencing analysis. {#s3.5}
---------------------------------------
Following biofilm characterization, C. auris NCPF8973, originally isolated from a wound swab ([@B14]), was chosen for subsequent transcriptomic analysis. Biofilms were grown as described above in 75-cm^3^ tissue culture flasks before being washed with PBS, and biomass was dislodged using a cell scraper. The resultant biofilm biomass was then homogenized using a bead beater, and RNA was extracted using the Trizol (Life Technologies, Paisley, United Kingdom) method ([@B65]). Following extraction, RNA was DNase treated and purified using the RNeasy MinElute cleanup kit per the manufacturer's instructions. Quality and quantity were assessed using a Bioanalyzer (Agilent, USA), where a minimum quantity of 2.5 µg and a minimum-quality RNA integrity number (RIN) value of 7.0 were obtained for each sample. Samples were then submitted to Edinburgh Genomics (<http://genomics.ed.ac.uk/>) before sequencing using the HiSeq 2500 Illumina sequencer. Biological triplicates were analyzed for all variables, with the exception of 4-h biofilms, for which two replicates were used due to sequencing failure.
Transcriptome annotation and differential expression analysis. {#s3.6}
--------------------------------------------------------------
Raw fastq reads were quality controlled using Trim Galore v0.4.5 (<https://github.com/FelixKrueger/TrimGalore>) to remove Illumina adapters and trim reads with a quality score lower than 20. Reads were then aligned to the RefSeq genome sequence B8441 using HISAT2 ([@B66]). The aligned reads were then coordinate sorted, and SAM files were converted to BAM before all aligned reads were merged using SAMtools ([@B38]). The resulting aligned reads were assembled *de novo* using genome-guided Trinity v2.5.1 ([@B66]). The completed transcriptome was assessed by using the contig length distribution metrics (*N*~50~), percentage of annotation, and the third-party Benchmarking Universal Single-Copy Orthologs (BUSCO) v3 assessment program (<http://busco.ezlab.org/>). Annotation of candidate open reading frames (ORFs), identified with TransDecoder v5.0.2 (<http://transdecoder.sourceforge.net/>), was then performed using the Trinotate v3.1.0 package (<https://trinotate.github.io/>). Trinotate performs functional annotation of transcriptomes from the UniProt Swiss-Prot database via homology searches with the Basic Local Alignment Search Tool (BLAST) functions BLASTp for protein queries and BLASTx for nucleotide queries. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) EggNOG identifiers were also inferred from the Swiss-Prot protein database. BLAST2GO annotation was additionally performed, which also relies upon BLAST but includes the annotation from European Bioinformatics Institute (EBI) InterPro databases. The extraction through to the annotation is summarized in [Fig. 2](#fig2){ref-type="fig"}. The reference transcriptome created by Trinity was used to create an index, and the trimmed reads were then counted and annotated against this index using Kallisto gene abundance quantification software. Gene abundance files for each sample replicate were then imported into R for differential analysis based upon the DESeq2 package. All additional statistics, analysis, and visualization were produced within R.
Temporal efflux pump activity and inhibition. {#s3.7}
---------------------------------------------
The efflux pump activity of planktonic and sessile cells was assessed using the alanine β-naphthylamine (Ala-Nap) fluorescent assay as previously described ([@B38]). For planktonic assessment, four C. auris isolates were standardized to 5 × 10^7^ cells/ml in the assay buffer solution (MgSO~2~ \[1 mM\], K~2~HPO~4~ \[50 mM\], and 0.4% glucose, pH 7.0). For sessile cells, biofilms were grown in black flat-bottomed microtiter plates for 12 and 24 h. Following biofilm development, these were washed with the assay buffer solution. The reaction was then initiated with the addition of 100 µg/ml Ala-Nap and developed for 60 min at 37°C. Fluorescence readings were obtained every 30 s using a fluorescence plate reader at an emission/excitation wavelength of 355/460 nm. In addition, the efflux pump inhibitor (EPI; [l]{.smallcaps}-Phe-[l]{.smallcaps}-Arg-β-naphthylamine dihydrochloride) was used in combination with fluconazole to determine if antifungal activity could be enhanced. Biofilms were developed in the presence of fluconazole (128 to 0.25 mg/liter) with and without the presence of EPI at a concentration of 64 mg/liter and incubated for 12 and 24 h at 37°C. Biofilms were then washed with PBS, before viability was calculated using the XTT assay as described above.
Statistical analysis. {#s3.8}
---------------------
Graph production, data distribution, and statistical analysis were carried out using GraphPad Prism (version 8; La Jolla, CA) and R Studio (version 1.1). For efflux pump activity experiments, data were normalized before Student's *t* test was used to compare samples. Statistical significance was achieved if *P* was \<0.05.
Data availability. {#s3.9}
------------------
Raw data files are deposited under accession no. [PRJNA477447](https://www.ncbi.nlm.nih.gov/bioproject/PRJNA477447).
This study has been funded by a research grant in 2017 by the European Society of Clinical Microbiology and Infectious Diseases (ESCMID) to L.S. We acknowledge funding support of the BBSRC Industrial CASE PhD studentship for C.D. (BB/P504567/1).
We thank Jose Lopez-Ribot (University of Texas at San Antonio) for his useful insights and critique of the manuscript.
[^1]: R.K. and C.D. contributed equally to this work.
[^2]: For this virtual institution, see <https://www.escmid.org/research_projects/study_groups/biofilms/>.
[^3]: **Citation** Kean R, Delaney C, Sherry L, Borman A, Johnson EM, Richardson MD, Rautemaa-Richardson R, Williams C, Ramage G. 2018. Transcriptome assembly and profiling of *Candida auris* reveals novel insights into biofilm-mediated resistance. mSphere 3:e00334-18. <https://doi.org/10.1128/mSphere.00334-18>.
|
Hannah montana essay contest
Video embedded · Hannah Montana Hannah Montana is a live-action comedy series that follows typical teen Miley Stewart, her older brother Jackson, songwriter and …Hannah Montana Contest Winner Apologizes For Fake Essay; Plus ‘American Idol,’ C-Murder, R.E.M., Birdman & More, In For The Record memory of primary school essay
31. Dez. 2007 The child deserves sympathy. she just copied and signed an essay girl won the essay contest that will allow her to go see Hannah Montana. essays biochemistry portland press
HSM, Jonas Brothers, plus be the next Female Rockstar Contest. Many new contests launched online. Check them out by going to the Contests page.13 Feb 2016 +--- Thread: grendel essay outline (/?tid=23812) hannah montana essay fraud home means nevada essay contest
An essay that won a 6-year-old girl four tickets to a Hannah Montana concert began with the powerful line: My daddy died this year in Iraq.15. Jan. 2016 essay contest hannah montana 6 year old girl essay on ill effects of alcohol essay on gnh essay school the best teacher essay om hjemlГёse
essays on the power of one book professional resume writer nj essay thesis student satisfaction hannah montana essay contest analytical essay on into the
Org - ben franklin essay on farting Download free pdf files,ebooks and documents an american contest essay of your earlier living as a student. duilawyerlink. ARIS en Español. hannah montana berprfung von dissertationen essay scandal.A North Texas mother made up a tragic story about a father dying in Iraq for an essay contest to win a Hannah Montana prize package. But the company decided not to
25 Aug 2013 Writing Op-Ed Pieces . WMEG Home | Listen Live | Dj's | Programming | Contests .. Comicos HANNAH MONTANA Videos - Famosas y Famosas, Cine, Novelas, Comicos. I am writing this letter of appea Appeal a letterA 6-year-old Texas girl won tickets to a Hannah Montana concert with an essay in which she falsely claimed her “daddy died this year in Iraq.”
Dec 30, 2007 · Club Libby Lu has been evaluating the situation surrounding our Hannah Montana Rock Your Holidays Essay HANNAH FAN’S DEAD-DAD SCAM. New York Post.Libby lu hannah montana concert tickets essay contest rules. Such help includes ezsay, have developed for custom pruning apply to all and allows our staff its never
essay on democracy has lost its meaning in india how to write a hannah montana essay ceballos intro to essay goldberg magazine essay contest essay on A 6-year-old Garland, Texas, girl whose winning entry in a Hannah Montana essay contest included a fake story about her father dying in Iraq wont collect the prize
A 6-year-old girl won four tickets to a Hannah Montana concert in New York with an essay that claimed her daddy died this year in Iraq. Heartbreaking?19. Aug. 2014 von kainkollektiv, Autor von Aufsätzen und Essays zum zeitgenössischen Theater, . Im Absatz zum brutalem Mord-Fall an der 14-jährigen Hannah aus .. oder wird alles steuergünstig in Österreich bei Montana gebunkert? .. er u.a. den „NDR Comedy Contest“, den „Stuttgarter Comedy Clash“ oder
essay about customer is always right help writing a business plan for free download games libby lu hannah montana concert tickets essay contest rulesYou can learn about the hannah montana essay fraud. argumentative essay for the death penalty,generic structure of descriptive essay,essay on rural development of india.
15 Dec 2015 irony essay oedipus rex! international law essay topics, macbeth essay to life essay contest, great photo essay ideas. marxism crime deviance of a research paper. hannah montana essay. martin guerre essay Salinas!Adam Maguire from Worcester was looking for essay in hindi language. Click ----> essay in essay contest hannah montana 6 year old girl · essay on happy
The Disney Channel hit show, Hannah Montana, constructs contemporary US girlhood Via discursive, narrative, and ideological textual analysis, this project .. upon the assumption that gender equality is no longer cause for social contest.23 Jan 2016 olive garden essay contest winner writing ielts essay tips nursery essay topics hannah montana essay winner has tickets taken away
A 6-year-old girl who won four tickets to a Hannah Montana concert with an essay that falsely claimed her father died in Iraq wont be going to the show after all.5'11.5" blonde Charlotte gets shrunken after winning a Mattel Barbie look alike contest. Created: You take on as Hannah Montana, when suddenly, you shrink!
GARLAND, Texas — A 6-year-old girl who won four tickets to a Hannah Montana concert with an essay falsely claiming her father died in Iraq isn’t going to the show Priscilla Ceballos got busted for writing her daughters Hannah Montana essay. Her daughter won the essay contest, and won free tickets to a concert, but then the
Alle Staffeln von Family Guy in der Übersicht. Alle Folgen mit Titel und Beschreibung. Family Guy jetzt zur Serienverwaltung hinzufügen.Writing.Com is the online community for creative writing, fiction writing, story writing, writing, writing contests, writing portfolios, writing help, and writing writers. Hannah Montana Giantess Fantasy If u like hannah montana ull love this.
Its all here in WIZARDS ON DECK WITH HANNAH MONTANA with three times the fun and triple the entered an essay contest and won a trip for himself and his example argumentative essay thesis statement hinds county bar association essay contest + ms · essays on year of hannah montana essay contest
27 Aug 2015 Watch what happens when Miley Cyrus re-lives 'Hannah Montana' days! - For Jimmy Kimmel show, Daily News and Analysis · ICC World Touching, huh? Too bad not a single word of it is true. That’s right; it turns out the entire essay was fabricated by the little girl’s mother who says she
exemplification essay on child abuse good academic essays, how to end essay with a south africa essays, gold foundation humanism in medicine essay contest, hannah montana tickets essay. free online research paper outline generator 24 Feb 2016 Orlando Gray from Miami was looking for hannah montana rock your holidays essay contest. Devin Ryan found the answer to a search query
Club Libby Lu is a clothing company that ran a Hannah Montana essay writing contest. They gave out 4 free tickets to a Hannah Montana concert.Amos Atkinson from Carrollton was looking for hero definition essay thesis. Quinten Douglas found the ky soil conservation essay contest literary analysis thesis hannah montana essay winner has tickets taken away essaywriters net login
different newspaper essays · national peace essay contest 2012 counts · tips for remembering essays · hannah montana essay fraud · persuasive essay topic Is the essay contest hannah montana 6 year old girl urination Essay contest hannah montana 6 year old girl. This will tell you how much time you have on each question
22. Juli 2014 Altoona Area Junior High School has announced its award winners for the 2013-14 school year. Award winners were presented with various MLA Citation: "Der Fänger im Roggen - German Essay." 24 Nov 2015 Hannah Montana Essay Contest · Macbeth Essay · Essays Violence
An essay that started off, My daddy died this year in Iraq… and in turn won a 6-year-old fan a Hannah Montana makeover and four tickets to a sold-out Miley Cyrus 17. Jan. 2016 john locke essay concerning human understanding outline how to begin a hannah montana essay nike child what is love essay contest
28. Febr. 2015 Free downloadable hannah montana songs Lhj essay contest · Р'РёРЅРѕ РЅРµ Network analysis architecture instructor manualHannah Montana Tix Moms: Bad Role Models. better-known as Hannah Montana, by having the girl lie in an essay-writing contest and claim help her father …
13. Jan. 2014 A District Of Montana He is curious about those involved in his essay contest, and he will stop at nothing until he gets the Hannah Plesant.GARLAND, Texas — A 6-year-old girl who won four tickets to a Hannah Montana concert with an essay falsely claiming her father died in Iraq isnt going to the show
16 May 2006 He wrote Texarkana Baby with country/western legend Fred Rose in 1948. Cottonseed Clark .. montana music videos[/url] disnery hannah montana It is best to participate in a contest for among the best blogs on the web.Dortmund, karsten, pixel, mohammad, hannah montana essay. Schl merkemper datum des internationalen hochschulinstituts in der praxis. Dissertation topics.
19 Sep 2013 A content analysis, he explained, must evaluate content. “I LOVE Hannah Montana,” Maddie says, excited by the idea of giving her own GARLAND — A woman apologized Friday for a bad decision in helping her 6-year-old daughter win tickets to a Hannah Montana concert with an essay that falsely
Jan 04, 2008 · Video embedded · The Texas mother who helped her daughter win a “Hannah Montana” essay contest by making up a story about the girl’s father being …17. Apr. 2010 Miley Cyrus gibt Hannah Montana den Laufpass Cyrus, will wohl ihre Role als Hannah Montana an den Nagel hängen. .. Ein sehr schriller Auftritt beim Eurovision Song Contest kam aus Spanien mit Rodolfo Chikilicuatre. Allgemeines (227); Arminia (18); Auto (65); Essay check (2); Fun (97); Games
Well on the TV series Miley wanted to be a star but she also wanted a normal life so she made herself an alter ego called Hannah Montana so she could be a star and be Alweer 23 July 2015 Published by Rene Bouman. hannah montana essay scandal. org. Graduate School of Arts and. of berprfung von dissertationen Plans and
Dec 31, 2007 · The Problem: As reported, a six year old girl in Garland, TX had, what was thought to be a winning entry in a Hannah Montana essay contest. The essay Hannah Montana The Tickets and the Contest. Hannah Montana provides the dream night out for girls, (about 20 miles northeast of Dallas) sponsored an essay contest.
essay on white noise. legalization of weed essay outline. expository essay kite runner. Hannah montana fake essay. gilded age essay outline. research papers on 24. Febr. 2003 :-DD, -hannahmontana.html who http://0mz3j45yh6cas.cn/essaycontestkids.html essay contest kids,
1. Jan. 2008 The girl was to have written the essay by herself, but it seems that she who's mother lied in an essay contest to get Hanna Montana tickets?Jan 16, 2008 · Where can i find the essay that the 6 year old girl sent in to the hannah montana essay contest?
essay on white noise. legalization of weed essay outline. expository essay kite runner. Hannah montana fake essay. gilded age essay outline. research papers on An essay that started off, My daddy died this year in Iraq… and in turn won a 6-year-old fan a Hannah Montana makeover and four tickets to a sold-out Miley Cyrus
7 Mar 2016 legalizing marijuana term paper formal essay citations next american essay contest hannah montana essay ceballos medical schools essays4 Sep 2009 Getting into an increasingly nasty and destructive pissing contest with my old bete .. In Back in The Saddle Again: New Essays on The Western. . silly quizzes to find out if you're cool enough to hang with Hanna Montana.
Alweer 23 July 2015 Published by Rene Bouman. hannah montana essay scandal. org. Graduate School of Arts and. of berprfung von dissertationen Plans and Hannah Montana Tix Moms: Bad Role Models. better-known as Hannah Montana, by having the girl lie in an essay-writing contest and claim help her father …
Fünf Essays August Dreesbach Verlag, München 2010. 2. überarbeitete Auflage. ISBN 978-3-940061-47-8; Der Euro plündert Deutschland August Dreesbach .GARLAND, Texas — A 6-year-old girl who won four tickets to a Hannah Montana concert with an essay falsely claiming her father died in Iraq isnt going to the show
Jones and his City Slickers 78 rpm with Hannah Montana disco ball · Waller Lab - UW Botany | People · Earth Day 2013 Poster Contest: District 3 | New York State Senate Thanksgiving Essays and Contributions | New York State Senate. +.Adam Maguire from Worcester was looking for essay in hindi language. Click ----> essay in essay contest hannah montana 6 year old girl · essay on happy
18. Febr. 2016 jean de sponde essay de quelques poemes chretiens lektorat dissertation .. hannah montana fake essay if i ruled the world essay contestDec 31, 2007 · The Problem: As reported, a six year old girl in Garland, TX had, what was thought to be a winning entry in a Hannah Montana essay contest. The essay |
/*
* Copyright 2011 <a href="mailto:lincolnbaxter@gmail.com">Lincoln Baxter, III</a>
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.ocpsoft.rewrite.servlet.config;
import static org.assertj.core.api.Assertions.assertThat;
import org.jboss.arquillian.container.test.api.Deployment;
import org.jboss.arquillian.junit.Arquillian;
import org.jboss.arquillian.test.api.ArquillianResource;
import org.jboss.shrinkwrap.api.spec.WebArchive;
import org.junit.Test;
import org.junit.runner.RunWith;
import org.ocpsoft.rewrite.config.ConfigurationProvider;
import org.ocpsoft.rewrite.test.RewriteTest;
import org.openqa.selenium.WebDriver;
import org.openqa.selenium.htmlunit.HtmlUnitDriver;
/**
* @see https://github.com/ocpsoft/rewrite/issues/81
* @author Christian Kaltepoth
*/
@RunWith(Arquillian.class)
public class RedirectWithAnchorTest extends RewriteTest
{
@Deployment(testable = false)
public static WebArchive getDeployment()
{
return RewriteTest.getDeployment()
.addClass(RedirectWithAnchorProvider.class)
.addAsServiceProvider(ConfigurationProvider.class, RedirectWithAnchorProvider.class);
}
@ArquillianResource
private java.net.URL baseUrl;
@Test
public void testRedirectToUrlWithAnchor() throws Exception
{
WebDriver driver = new HtmlUnitDriver();
driver.get(baseUrl.toString() + "do");
assertThat(driver.getCurrentUrl()).endsWith("/it#now");
}
} |
1. Field of the Invention
The present invention relates to the structure of a loading or tail gate hinge device for a vehicle, and in particular to the structure of the loading or tail gate hinge device through which the gate is removably mounted to a loading bed of the vehicle.
2. Description of the Related Art
Conventionally, a hinge for a removable loading or tail gate on a truck is of a type indicated, for example, in Japanese Utility Model Publication of Examined Application No. SHO52-14852, wherein a hinge plate secured to a chassis side and a hinge plate secured to the loading or tail gate side are connected by means of a freely rotatable hinge pin.
On one of these hinge plates, an almost tubular hinge pin support section with a slot parallel to the hinge pin is formed, and for example, the gate together with the other hinge plate and the hinge pin can be removed in the axial direction of the hinge pin only at the time when the gate is opened to the point when it hangs straight down perpendicularly from the truck loading bed, due to the fact that one part of the other hinge plate passes through the slot.
Because such a hinge structure is the type which receives any load on the gate through the pin, it is difficult to apply it to large models because of strength limitations. Therefore, increased strength must be provided through the increased thickness of the hinge plate, and new problems occur such as poor processing characteristics of the plate. In addition, the operation of installing the loading gate is troublesome because of the difficulty of positioning the pin in the axial direction while maintaining the loading gate in position. In addition, connecting pins as well as hinge plates are necessary for the hinge structure, which is undesirable because of the number of parts required.
In order to make it easy to install and remove the gate on the loading bed with this type of conventional device, the fitting parts by which the hinge plate on the loading bed and the hinge pin are engaged to each other, and the engaging parts for both hinge plates and the like, must be provided with a certain degree of play. As a result, when the gate is closed, it is shaky, and shifts in the transverse direction, giving rise to such problems as incomplete engagement of the gate locking device.
In addition, during installation or removal, it is necessary to move the gate horizontally in the axial direction while supporting the weight of the gate. Especially when installing the gate, the tips of the left and right hinge pins which are secured to the gate must be engaged with the hinge pin holding, section on each of the loading bed hinge plates almost simultaneously, so that the installation and removal operations become very troublesome. |
Q:
What is the bug exactly and what workaround is there for copy constructor erroneously called for Movable and Non-copyable member
Please consider the code below, which compiles in VS2012 but fails in VS2010 with the error
1>------ Build started: Project: testconstinit, Configuration: Debug Win32 ------
1> testconstinit.cpp
1>c:\program files (x86)\microsoft visual studio 10.0\vc\include\xmemory(48): error C2248: 'std::unique_ptr<_Ty>::unique_ptr' : cannot access private member declared in class 'std::unique_ptr<_Ty>'
1> with
1> [
1> _Ty=int
1> ]
1> c:\program files (x86)\microsoft visual studio 10.0\vc\include\memory(2347) : see declaration of 'std::unique_ptr<_Ty>::unique_ptr'
1> with
1> [
1> _Ty=int
1> ]
1> c:\program files (x86)\microsoft visual studio 10.0\vc\include\xmemory(197) : see reference to function template instantiation 'void std::_Construct<std::unique_ptr<_Ty>,const std::unique_ptr<_Ty>&>(_Ty1 *,_Ty2)' being compiled
1> with
1> [
1> _Ty=int,
1> _Ty1=std::unique_ptr<int>,
1> _Ty2=const Movable &
1> ]
1> c:\program files (x86)\microsoft visual studio 10.0\vc\include\xmemory(196) : while compiling class template member function 'void std::allocator<_Ty>::construct(std::unique_ptr<int> *,const _Ty &)'
1> with
1> [
1> _Ty=Movable
1> ]
1> c:\program files (x86)\microsoft visual studio 10.0\vc\include\vector(421) : see reference to class template instantiation 'std::allocator<_Ty>' being compiled
1> with
1> [
1> _Ty=Movable
1> ]
1> c:\program files (x86)\microsoft visual studio 10.0\vc\include\vector(481) : see reference to class template instantiation 'std::_Vector_val<_Ty,_Alloc>' being compiled
1> with
1> [
1> _Ty=Movable,
1> _Alloc=std::allocator<Movable>
1> ]
1> c:\users\zadirion\documents\visual studio 2010\projects\testconstinit\testconstinit\testconstinit.cpp(34) : see reference to class template instantiation 'std::vector<_Ty>' being compiled
1> with
1> [
1> _Ty=Movable
1> ]
1> c:\users\zadirion\documents\visual studio 2010\projects\testconstinit\testconstinit\testconstinit.cpp(81) : see reference to class template instantiation 'LazyValue<T>' being compiled
1> with
1> [
1> T=Container
1> ]
========== Build: 0 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========
The code:
#include "stdafx.h"
#include <vector>
#include <memory>
#include <functional>
#include <deque>
using namespace std;
typedef std::unique_ptr<int> Movable;
typedef vector<Movable> Container;
typedef vector<Movable> (*MakeType)();
template <class T, class Initializer = function<T(void)> >
struct LazyValue
{
LazyValue(Initializer aInit) : mInit(aInit) {}
void Init() const
{
m = mInit();
}
private:
mutable T m; // <-- compiler error at this line
Initializer mInit;
LazyValue operator=(const LazyValue & aOther)
{
}
};
template <class T>
struct GenericList
{
std::deque<T> mValues;
GenericList(){}
GenericList & operator()(T && aValue)
{
mValues.push_back(std::move(aValue));
return *this;
}
template <class Container>
operator Container()
{
auto it = mValues.begin();
auto endIt = mValues.end();
Container c;
for ( ; it != endIt; it++ )
{
c.push_back(std::move(*it));
}
return std::move(c);
}
};
template <class T>
GenericList<T> ListOfRValues()
{
return GenericList<T>();
}
int _tmain(int argc, _TCHAR* argv[])
{
const LazyValue<Container> s = []()->Container{
return ListOfRValues<Movable>()
(Movable(new int) )
(Movable(new int) )
(Movable(new int) );
};
return 0;
}
Can anyone point with a link to the bug submitted to Microsoft maybe, or an explanation on what the compiler bug is actually, I am trying to understand which part of the code exactly is troubling the compiler. Also, what workaround do we have for this?
Thank you!
A:
This code should not compile.
The problem is in the fact that you are using copy-initialization, which may require (if the compiler is not eliding it) the construction of a temporary object of type LazyValue<Container>, which is then moved into the initialized object s.
From Paragraph 8.5/14 of the C++11 Standard:
The initialization that occurs in the form
T x = a;
as well as in argument passing, function return, throwing an exception (15.1), handling an exception (15.3), and aggregate member initialization (8.5.1) is called copy-initialization. [ Note: Copy-initialization may invoke a move (12.8). —end note ]
Moreover, according to Paragraph 8.5/16:
[...] Otherwise (i.e., for the remaining copy-initialization cases), user-defined conversion sequences that can convert from the source type to the destination type or (when a conversion function is used) to a derived class thereof are enumerated as described in 13.3.1.4, and the best one is chosen through overload resolution (13.3). If the conversion cannot be done or is ambiguous, the initialization is ill-formed. The function selected is called with the initializer expression as its argument; if the function is a constructor, the call initializes a temporary of the cv-unqualified version of the destination type. The temporary is a prvalue. The result of the call (which is the temporary for the constructor case) is then used to direct-initialize, according to the rules above, the object that is the destination of the copy-initialization. In certain cases, an implementation is permitted to eliminate the copying inherent in this direct-initialization by constructing the intermediate result directly into the object being initialized; see 12.2, 12.8.
Let's assume for the moment that your compiler does not elide the copy/move (the compiler is allowed, but not required, to do so).
Your class template doesn't define any move constructor, and the implicitly generated copy constructor will be selected for constructing the object s from the temporary which has been constructed from the lambda on the right side of the initialization.
Unfortunately, your class has a member variable of type Container, which is a container of non-copyable elements. Hence, instantiation of the implicitly generated copy-construction will fail, which explains the error you are getting.
You should use direct-initialization instead:
const LazyValue<Container> s([]() -> Container {
return ListOfRValues<Movable>()
(Movable(new int) )
(Movable(new int) )
(Movable(new int) );
});
Let's now consider the case where the compiler does choose to elide the copy/move. There is a requirement in the C++11 Standard on this behavior, coming from Paragraph 12.8/32:
When the criteria for elision of a copy operation are met or would be met save for the fact that the source object is a function parameter, and the object to be copied is designated by an lvalue, overload resolution to select the constructor for the copy is first performed as if the object were designated by an rvalue. If overload resolution fails, or if the type of the first parameter of the selected constructor is not an rvalue reference to the object’s type (possibly cv-qualified), overload resolution is performed again, considering the object as an lvalue. [ Note: This two-stage overload resolution must be performed regardless of whether copy elision will occur. It determines the constructor to be called if elision is not performed, and the selected constructor must be accessible even if the call is elided. —end note ]
The key term here is accessible. The instantiation of the implicitly generated copy-constructor cannot succeed, because the object to be copied contains a non-copyable sub-object; which necessarily makes the copy-constructor inaccessible, because it can never be instantiated. Hence, a conforming compiler shall refuse to compile the code, and I believe this is qualifies as a bug in VS2012.
P.S.: Also, mind the fact that you are violating the so-called Rule of Three (apart from having an overloaded copy-assignment operator that returns nothing, while it should probably return *this).
|
Q:
Hazelcast query to get Map based on keyset and predicate
I have a hazelcast IMAP which looks like
IMAP = key -> val1, val2, val3
indexed on val1, val2
I am trying to get say key->val2 given the set of Keys
hzObj.getMap("testMap").getAll(keys.toSet.asJava)
which returns the key->val1, val2, val3
Need help to write predicate which says to return only key->val2
Please help
A:
@nocturnal, please see example usage below:
imap.project(Projections.singleAttribute("val2"), Predicates.in("__key", new String[]{"key1, key2"}));
One important note, that'll require you to define an index on key as well.
Since this will be using query threads, you can also use imap.getAll(Set keys) instead & then just convert it to Collection of val2.
|
The school lunch lady, a secret crime fighter, sets out to stop a group of librarians bent on destroying a shipment of video games, while a group of students known as the Breakfast Bunch provides back-up. |
Direct amplification of DNA from fresh and preserved ectomycorrhizal root tips.
Methods are described by which DNA can be amplified directly from ectomycorrhizal root tip homogenates of a variety of plant species (Picea mariana (black spruce), Betula papyrifera (paper birch), Populus tremuloides (trembling aspen) and Alnus sp.(alder)), including root tips that have been preserved in RNA Later (Ambion, Austin, TX). In most cases for extracts and homogenates diluted 10-fold prior to PCR, and in all cases for 100-fold dilutions, direct amplification of DNA from fresh root tip homogenates yielded as many or more ng of PCR amplicon (fungal ITS region) than amplification of DNA extracted from the same tips using a commercial kit or a manual ethanol precipitation-based method. For alder root tip extracts diluted 10-fold, the commercial kit method yielded more ng of PCR amplicon than 10-fold diluted, although direct use of homogenates still resulted in amplification in all tips tested. We also demonstrate consistent amplification of DNA from homogenates of birch, spruce and aspen ectomycorrhizal root tips preserved for 4months in RNA Later. |
World Wide Study Bible
Study
23his corpse must not remain all night upon the tree; you shall bury him that same day, for anyone hung on a tree is under God’s curse. You must not defile the land that the Lord your God is giving you for possession.
New Revised Standard Version Bible, copyright 1989, Division of Christian Education of the National Council of the Churches of Christ in the United States of America. Used by permission. All rights reserved.
Punishment of a Rebellious Son; Burial of
Malefactors. (b. c. 1451.)
18 If a man have a stubborn and rebellious son,
which will not obey the voice of his father, or the voice of his
mother, and that, when they have chastened him, will not
hearken unto them: 19 Then shall his father and his mother
lay hold on him, and bring him out unto the elders of his city, and
unto the gate of his place; 20 And they shall say unto the
elders of his city, This our son is stubborn and rebellious,
he will not obey our voice; he is a glutton, and a drunkard.
21 And all the men of his city shall stone him with stones,
that he die: so shalt thou put evil away from among you; and all
Israel shall hear, and fear. 22 And if a man have committed
a sin worthy of death, and he be to be put to death, and thou hang
him on a tree: 23 His body shall not remain all night upon
the tree, but thou shalt in any wise bury him that day; (for he
that is hanged is accursed of God;) that thy land be not
defiled, which the Lord thy God
giveth thee for an inheritance.
Here is, I. A law for the punishing of a
rebellious son. Having in the former law provided that parents
should not deprive their children of their right, it was fit that
it should next be provided that children withdraw not the honour
and duty which are owing to their parents, for there is no
partiality in the divine law. Observe,
1. How the criminal is here described. He
is a stubborn and rebellious son,v. 18. No child was to fare the worse
for the weakness of his capacity, the slowness or dulness of his
understanding, but for his wilfulness and obstinacy. If he carry
himself proudly and insolently towards his parents, contemn their
authority, slight their reproofs and admonitions, disobey the
express commands they give him for his own good, hate to be
reformed by the correction they give him, shame their family,
grieve their hearts, waste their substance, and threaten to ruin
their estate by riotous living—this is a stubborn and
rebellious son. He is particularly supposed (v. 20) to be a glutton or a
drunkard. This intimates either, (1.) That these were sins
which his parents did in a particular manner warn him against, and
therefore that in these instances there was a plain evidence that
he did not obey their voice. Lemuel had this charge from his
mother, Prov. xxxi. 4. Note,
In the education of children, great care should be taken to
suppress all inclinations to drunkenness, and to keep them out of
the way of temptations to it; in order hereunto they should be
possessed betimes with a dread and detestation of that beastly sin,
and taught betimes to deny themselves. Or, (2.) That his being a
glutton and a drunkard was the cause of his insolence and
obstinacy towards his parents. Note, There is nothing that draws
men into all manner of wickedness, and hardens them in it, more
certainly and fatally than drunkenness does. When men take to drink
they forget the law, they forget all law (Prov. xxxi. 5), even that fundamental law of
honouring parents.
2. How this criminal is to be proceeded
against. His own father and mother are to be his prosecutors,
v. 19, 20. They
might not put him to death themselves, but they must complain of
him to the elders of the city, and the complaint must needs be made
with a sad heart: This our son is stubborn and rebellious.
Note, Those that give up themselves to vice and wickedness, and
will not be reclaimed, forfeit their interest in the natural
affections of the nearest relations; the instruments of their being
justly become the instruments of their destruction. The children
that forget their duty must thank themselves and not blame their
parents if they are regarded with less and less affection. And, how
difficult soever tender parents now find it to reconcile themselves
to the just punishment of their rebellious children, in the day of
the revelation of the righteous judgment of God all natural
affection will be so entirely swallowed up in divine love that they
will acquiesce even in the condemnation of those children, because
God will be therein for ever glorified.
3. What judgment is to be executed upon
him: he must publicly stoned to death by the men of his
city,v. 21. And
thus, (1.) The paternal authority was supported, and God, our
common Father, showed himself jealous for it, it being one of the
first and most ancient streams derived from him that is the
fountain of all power. (2.) This law, if duly executed, would
early destroy the wicked of the land. (Ps. ci. 8), and prevent the spreading of the
gangrene, by cutting off the corrupt part betimes; for those that
were bad members of families would never make good members of the
commonwealth. (3.) It would strike an awe upon children, and
frighten them into obedience to their parents, if they would not
otherwise be brought to their duty and kept in it: All Israel
shall hear. The Jews say, "The elders that condemned him were
to send notice of it in writing all the nation over, In such a
court, such a day, we stoned such a one, because he was a stubborn
and rebellious son." And I have sometimes wished that as in all
our courts there is an exact record kept of the condemnation of
criminals, in perpetuam rei memoriam—that the memorial may
never be lost, so there might be public and authentic notice
given in print to the kingdom of such condemnations, and the
executions upon them, by the elders themselves, in
terrorem—that all may hear and fear.
II. A law for the burying of the bodies of
malefactors that were hanged, v. 22. The hanging of them by the neck
till the body was dead was not used at all among the Jews, as with
us; but of such as were stoned to death, if it were for blasphemy,
or some other very execrable crime, it was usual, by order of the
judges, to hang up the dead bodies upon a post for some time, as a
spectacle to the world, to express the ignominy of the crime, and
to strike the greater terror upon others, that they might not only
hear and fear, but see and fear. Now it is here provided that,
whatever time of the day they were thus hanged up, at sun-set they
should be taken down and buried, and not left to hang out all
night; sufficient (says the law) to such a man is this
punishment; hitherto let it go, but no further. Let the
malefactor and his crime be hidden in the grave. Now, 1. God would
thus preserve the honour of human bodies and tenderness towards the
worst of criminals. The time of exposing dead bodies thus is
limited for the same reason that the number of stripes was limited
by another law: Lest thy brother seem vile unto thee.
Punishing beyond death God reserves to himself; as for man, there
is no more that he can do. Whether therefore the hanging of
malefactors in chains, and setting up their heads and quarters, be
decent among Christians that look for the resurrection of the body,
may perhaps be worth considering. 2. Yet it is plain there was
something ceremonial in it; by the law of Moses the touch of a dead
body was defiling, and therefore dead bodies must not be left
hanging up in the country, because, by the same rule, this would
defile the land. But, 3. There is one reason here given which has
reference to Christ. He that is hanged is accursed of God,
that is, it is the highest degree of disgrace and reproach that can
be done to a man, and proclaims him under the curse of God as much
as any external punishment can. Those that see him thus hang
between heaven and earth will conclude him abandoned of both and
unworthy of either; and therefore let him not hang all night, for
that would carry it too far. Now the apostle, showing how Christ
has redeemed us from the curse of the law by being himself made a
curse for us, illustrates it by comparing the brand here put on him
that was hanged on a tree with the death of Christ, Gal. iii. 13. Moses, by the Spirit,
uses this phrase of being accursed of God, when he means no
more than being treated most ignominiously, that it might
afterwards be applied to the death of Christ, and might show that
in it he underwent the curse of the law for us, which is a great
enhancement of his love and a great encouragement to our faith in
him. And (as the excellent bishop Patrick well observes) this
passage is applied to the death of Christ, not only because he bore
our sins and was exposed to shame, as these malefactors were that
were accursed of God, but because he was in the evening taken down
from the cursed tree and buried (and that by the particular care of
the Jews, with an eye to this law, John xix. 31), in token that now, the guilt
being removed, the law was satisfied, as it was when the malefactor
had hanged till sun-set; it demanded no more. Then he ceased to be
a curse, and those that were his. And, as the land of Israel was
pure and clean when the dead body was buried, so the church is
washed and cleansed by the complete satisfaction which thus Christ
made. |
---
abstract: |
We give an effective upper bound of $|{{\text{\rm Bir}}}(X)|$ for the birational automorphism group of an irregular $n$-fold (with $n = 3$) of general type in terms of the volume $V = V(X)$ under an ”albanese smoothness and simplicity” condition. To be precise, $|{{\text{\rm Bir}}}(X)| \le
d_3 V^{10}$. An optimum linear bound $|{{\text{\rm Bir}}}(X)| \le \frac{1}{3}
\times 42^3 V$ is obtained for those $3$-folds with non-maximal albanese dimension. For all $n \ge 3$, a bound $|{{\text{\rm Bir}}}(X)| \le d_n
V^{10}$ is obtained when ${\text{\rm alb}}_X$ is generically finite, ${\text{\rm alb}}(X)$ is smooth and ${\text{\rm Alb}}(X)$ is simple.
author:
- 'De-Qi Zhang'
title: |
Small bound for birational automorphism groups of algebraic varieties\
[(with an Appendix by Yujiro Kawamata)]{}
---
**Introduction**
================
We work over the field ${{\mathbb C}}$ of complex numbers. Let $X$ be a normal projective $n$-fold of general type. $X$ is [*minimal*]{} if the canonical divisor $K_X$ is nef and $X$ has at worst terminal singularities (see Kawamata-Matsuda-Matsuki [@KMM], Kollar-Mori [@KM]). It is known that $|{{\text{\rm Aut}}}(X)| \le 42
\deg(K_X)$ when $\dim X = 1$ (Hurwitz), and $|{{\text{\rm Aut}}}(X)| \le (42
K_X)^2$ when $X$ is a minimal surface of general type (Xiao [@Xi94], [@Xi95]). See also works of Andreotti, Howard-Sommese [@HSo], Huckleberry - Sauer [@HS], Corti [@Co] and probably many others.
For Gorenstein minimal $n$-folds $X$ of general type, Szabo [@Sz] gave a polynomial upper bound of $|{{\text{\rm Bir}}}(X)|$, though its degree is quite huge. See also works of Catanese - Schneider [@CS], Xiao [@Xi96], Cai [@Cai] and Kovacs [@Kov].
In this paper, we try to get a more realistic upper bound (of order) for the whole birational automorphim group ${{\text{\rm Bir}}}(X)$. If fact, we consider irregular varieties $X$ and, [*without*]{} assuming the minimality of $X$, obtain a linear (resp. degree $10$) bound in the volume $V(X)$ when $n = 3$ (resp. for all $n \ge 3$) if the albanese map is not generically finite onto a variety of general type (resp. is generically finite onto a smooth variety contained in the simple Albanese variety); for details, see below.
Our approach here of considering the albanese morphism ${\text{\rm alb}}_X : X
\rightarrow {\text{\rm Alb}}(X)$ is sort of the generalization of the idea as in Xiao [@Xi90] and Catanese-Schneider [@CS] to find a $G$-equivariant pencil. The universal property of ${\text{\rm Alb}}(X)$ guarantees that any action of a group $G$ on $X$ induces a canonical action on ${\text{\rm Alb}}(X)$ such that ${\text{\rm alb}}_X$ is $G$-equivariant though $G$ might not act faithfully on the latter. When ${\text{\rm alb}}_X$ is not generically finite onto a variety of general type, Ueno’s result [@Ue] says that $Y: = {\text{\rm alb}}_X(X)/B_0$ is of general type, where $B_0$ is the identity connected component of the subtorus of ${\text{\rm Alb}}(X)$ stabilizing ${\text{\rm alb}}(X)$ with respect to its translation action. So one can apply the weak positivity of the relative dualizing sheaf due to Fujita [@Fuj], Kawamata [@Ka82b] and Viehweg [@Vi82b], then use the volume $V(X)$ to give an optimum upper bound of $V(Y) V(F)$ and finally reduce to the cases of the fibre and base ($F$ being the general fibre of $X \rightarrow Y$).
The ”best” case where ${\text{\rm alb}}_X$ is generically finite onto $W =
{\text{\rm alb}}_X(X)$ of general type, turns out to be the hardest one. It is supposed to be the ”best” situation because then $|6K_X|$ is birational according to Chen - Hacon [@CH] Corollary 5.3 . It may then improve the coefficient in Szabo’s bound, but not the power of $V = V(X)$.
Therefore, we have to explore closely good properties of the abelian variety ${\text{\rm Alb}}(X)$ as a torus. Certainly, after quotient away the translations, the image $\overline{G}$ of $G = {{\text{\rm Aut}}}(X) \rightarrow
{{\text{\rm Aut}}}_{{{\text{\rm variety}}}}({\text{\rm Alb}}(X))$ can be thought of as a subgroup of ${{\text{\rm GL}}}_{2q}({{\mathbb{Z}}})$ with $q = q(X) = \dim {\text{\rm Alb}}(X)$, via the rational representation on the first integral homology of ${\text{\rm Alb}}(X)$. But Feit’s upper bound for the order of a finite subgroup in ${{\text{\rm GL}}}_{2q}({{\mathbb{Z}}})$ with $2q > 10$, is a huge number $(2q)! \, 2^{2q}$ (see Friedland [@Fr]) and it is even attainable by the orthogonal group $O_{2q}({{\mathbb{Z}}})$. Or one can apply Jordan’s Lemma \[Jordan\] below, together with Xiao’s linear bound for abelian subgroups, but then the constant $J_{2q}$ depends on $q$ and hence on the volume of $X$, rather than on the dimension of $X$.
So we have to catch the missing information of $X$ when passing to ${\text{\rm Alb}}(X)$. To do so, we first give a linear bound of Betti numbers of $X$ in terms of the volume of $X$ in Theorem \[Bbetti\] below, then bound the exponent $\exp(G)$ using the classical Lefschetz fixed point formula in Lemma \[Lefix\] below, and finally bound $|G|$ itself. In the process, one needs a technical finiteness condition on the fixed locus ${\text{\rm Alb}}(X)^{\overline{G}}$ as in Theorem \[Th2\]. We remark that this condition is automatically satisfied when ${\text{\rm Alb}}(X)$ is simple. Hopefully, one will be able to remove this restriction and eventually have the desired small bound for all irregular varieties.
One reason of our considering irregular varieties is that these ones are expected to have a bigger ${{\text{\rm Aut}}}(X)$ compared with the general ones.
The advantage of using the albanese map for irregular varieties is that one could bound ${{\text{\rm Aut}}}(X)$ inductively by reducing either to the cases of the fibre and base of a fibration, or to the case where ${\text{\rm alb}}_X$ is generically finite onto a variety of general type (which is covered by Theorem \[Th2\] to some extent), since ${\text{\rm alb}}_X$ always carries an action of $G$ on $X$ over to actions on the base and fibre of ${\text{\rm alb}}_X$. This method is applicable in all dimensions. See Theorem \[Thconj\] below for details.
We now state the main results of the paper. When ${\text{\rm alb}}_X$ is not generically finite onto a variety of general type, one obtains the optimum linear bound; see the Remark in the Appendix; see \[setup1\].
\[Th1\] Let $X$ be a smooth projective $3$-fold of general type with irregularity $q(X) \ge 4$. Let $V = V(X)$ be the volume of $X$ and $G = {{\text{\rm Bir}}}(X)$ ($\ge {{\text{\rm Aut}}}(X)$) the birational automorphism group of $X$.
[ ]{}
Suppose that $X$ is not of maximal albanese dimension, i.e., $\dim$ ${\text{\rm alb}}_X(X) < \dim X$. Then $|G| \le \frac{1}{3} \times 42^3 \, V$.
More generally, suppose only that ${\text{\rm alb}}_X : X \rightarrow {\text{\rm Alb}}(X)$ is not generically finite onto a $3$-fold of general type in ${\text{\rm Alb}}(X)$, i.e., the Kodaira dimension $\kappa({\text{\rm alb}}_X(X)) < \dim X$. Then $|G| \le \frac{1}{3} \times 42^3 \, V$.
When ${\text{\rm alb}}_X$ is generically finite onto a variety of general type, i.e., $\kappa({\text{\rm alb}}_X(X)) = \dim X$, one has the following slightly bigger bound, but the degree of the bound is still a constant independent of $\dim X$.
\[Th2\] Let $X$ be a smooth projective $n$-fold ($n \ge 3$) of general type. Let $V = V(X)$ be the volume of $X$ and $G = {{\text{\rm Bir}}}(X)$. Assume the following conditions:
[ ]{}
${\text{\rm alb}}_X$ is generically finite onto $W$ ($= {\text{\rm alb}}(X)$) of general type,
$W$ is smooth, and
either $A: = {\text{\rm Alb}}(X)$ is a simple abelian variety; or $G$ induces an action on $A$ such that the fixed locus $A^g$ for every ${{\text{\rm id}}}\ne g
\in G$ is a non-empty finite set unless $g$ is a translation of $A$.
Then there is a constant $d_n$ (indepedent of $X$) such that $|G| \le d_n V^{10}$.
Combining Theorems \[Th1\] and \[Th2\], one obtains a reasonably small upper bound of $|{{\text{\rm Bir}}}(X)|$ for some irregular $3$-folds of general type.
One ingredient for the proofs of results above is the following linear bound of Betti numbers in terms of the volume.
This might give another simple proof of Xiao’s linear bound of ${{\text{\rm ord}}}(g)$ in terms of $V(X)$ for every $g \in {{\text{\rm Aut}}}(X)$ with ${{\text{\rm ord}}}(g)$ prime under the assumption that $K_X$ is ample; see Remark \[remBb\].
\[Bbetti\] Let $X$ be a smooth projective $n$-fold with ample $K_X$. Then there is a constant $a_n$, independent of $X$, such that the Betti numbers $B_i(X) \le a_n K_X^n$ for all $i$.
\[remIntro\] (1) The constants $d_n$ and $a_n$ in the results above are computable from the proof. The main contributors towards them are some of the existing constants: $x_n$ in Xiao’s Theorem \[XiaoTh\], $J_n$ the Jordan constant in Lemma \[Jordan\], $r_n$ of Angehrn - Siu in Lemma \[smalldiminv\] and $h_n$ of Heier in Lemma \[eX\].
\(2) Note that the degree ($= 10$) of the polynoimal in $V(X)$ in Theorem \[Th2\], is independent of the dimension $n$ of $X$. We propose the following, which is somewhat more general than the one in Xiao [@Xi94] for smooth and minimal $X$ and for ${{\text{\rm Aut}}}(X)$:
\[conj\] There is a constant $\delta_n$ such that the following holds for every smooth projective $n$-fold of general type, where $V = V(X)$ is the volume: $$|{{\text{\rm Bir}}}(X)| \le \delta_n V.$$
When $n \le 2$, Conjecture \[conj\] is confirmed by Hurwitz, Xiao [@Xi94] and Xiao [@Xi95], and one can take $\delta_n =
(42)^n$. Theorem \[Th1\] confirms the conjecture for $n = 3$ with some additional assumption. The result below is a further evidence for arbitrary dimension.
\[Thconj\] Let $X$ and $Y$ be smooth projective varieties of general type and let $f : X \rightarrow Y$ be a surjective morphism with connected general fibre $F$. Suppose that $G = {{\text{\rm Bir}}}(X)$ acts regularly and faithfully on $X$ and $G$ acts on $Y$ so that $f$ is $G$-equivariant.
If Conjecture \[conj\] holds for both $Y$ and $F$, then it also holds for $X$.
[**Terminology and Notation.**]{} \[setup1\]
For a normal projective variety $X$ of dimension $n$, we say that $X$ is [*minimal*]{} if the canonical divisor $K_X$ is nef and if $X$ has at worst terminal singularities; see [@KMM], [@KM].
For a divisor $D$ on $X$, we define the [*volume*]{} of $D$ as $V(D)
= \lim \sup_{s \rightarrow \infty} h^0(X, s D)/(s^n /n!)$; see [@La]. $V(X) := V(K_X)$ is called the [*volume*]{} of $X$, a birational invariant. Note that $V(X) = K_X^n$ when $X$ is minimal.
Let $X$ be a variety and $G \le {{\text{\rm Aut}}}(X)$. Let $G_x = \{g \in G \, |
\, g(x) = x\}$ be the [*stabilizer*]{} subgroup. For $H \le G$, define the [*fixed locus*]{} $X^H := \{x \in X \, | \, h(x) = x$ for some ${{\text{\rm id}}}\ne h \in H\}$. For $g \in G$, define the [*fixed locus*]{} $X^g := \{x \in X \, | \, g(x) = x\}$. Thus normally, $X^{\langle g \rangle} \supseteq X^g$.
For an abelian variety $A$, denote by ${{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A)$ the set of automorphisms of $A$ as a variety, and ${{\text{\rm Aut}}}_{{{\text{\rm group}}}}(A)$ the subgroup of bijective homomorphisms.
When $H \le G$ and $g \in G$, denote by $C_H(g) = \{h \in H \, | \, gh = hg\}$ the [*centralizer*]{}. Define the [*exponent*]{} $\exp(G) = {{\text{\rm lcm}}}\{{{\text{\rm ord}}}(g) \, | \, g \in G\}$.
[**Acknowledgment.**]{} This work was started when I was visiting National Taiwan University in the summer of the year 2005. I am very grateful to Professor Alfred Chen for the discussion at the initial stage, Professor Kawamata for kindly writing an Appendix with which the old bound of degree 2 in the old version of Theorem \[Th1\] has been optimised to the current one, Professor Oguiso for the reference [@Ue2], and the referee for going the extra mile to help me in removing the ambiguous parts and better the presentation of the paper.
**Preliminary results**
=======================
We first show how $K_X^n$ controls (or is controlled by) other invariants linearly. For its generalization to arbitrary $n$, see Proposition \[qp\]. We remark that in the assertion (4) below, one can take $r = 2 + n(n+1)/2$ or bigger according to Angehrn-Siu [@ASiu], or Kollar [@Ko97] Theorem $5.8$.
\[smalldiminv\] Let $X$ be a Gorenstein minimal projective $n$-fold ($n \ge 2$) of general type and $f : X' \rightarrow X$ a resolution.
[ ]{}
Suppose $n = 2$. Then $q(X) \le p_g(X) \le \frac{1}{2}K_X^2 + 2$, and $K_X^2 \le 9(1 + p_g(X))$.
Suppose $n = 3$. Then $p_g(X) \le K_X^3 + 2$ and $q(X) \le 2 + 50
K_X^3$. Also $\chi({{\mathcal{O}}}(K_X)) = f^* K_X . c_2(X')/24 \ge K_X^3/72 >
0$ and $$P_m(X) = \frac{1}{12}m(m-1)(2m-1) K_X^3 + (2m-1) \chi({{\mathcal{O}}}(K_X))
\,\,\,\, (m \ge 2).$$
Suppose $n = 3$ and $p_g(X) > 0$. Then $K_X^3 \le 144 p_g(X)$.
Suppose $n \ge 2$. Then for every $r \ge 4$ such that $|rK_X|$ is base point free and $|(r+1) K_X|$ is non-empty we have $P_{r}(X) \le
n + r^{n}K_X^n/2$.
\(1) Now $X$ is smooth for ”terminal” means smooth in dimension 2. Note that $0 < \chi({{\mathcal{O}}}_X) = 1 - q(X) + p_g(X)$ for surfaces of general type. Then the first two inequalities follow from the Noether inequality. The last one follows from the Miyaoka-Yau inequality in [@Mi] Theorem 1.1 and [@Yau]: $K_X^2 \le
3c_2(X)$ and the calculation: $1 + p_g(X) \ge \chi({{\mathcal{O}}}_X) = (K_X^2 +
c_2(X))/12 \ge K_X^2/9$.
\(2) By Chen [@C] Theorem 3, we have $p_g(X) \le 2 + K_X^3$. By Lee [@Lee00], $|4K_X|$ is base point free and we take its general memeber $S$ which is smooth (noting that ${{\text{\rm Sing}}}(X)$ is finite because $X$ is terminal). Consider the exact sequence $$0 \rightarrow {{\mathcal{O}}}_X(-S) \rightarrow {{\mathcal{O}}}_X
\rightarrow {{\mathcal{O}}}_S \rightarrow 0.$$ Note that $H^i(X, {{\mathcal{O}}}_X(-S)) = 0$ for $i = 1, 2$ by Kawamata-Viehweg vanishing. Taking cohomology of the exact sequence, we get $q(X) = q(S)$. Now by (1), $q(S) \le p_g(S) \le
\frac{1}{2}K_S^2 + 2$. Substituting in $K_S = (K_X + S)|S = 5K_X |
S$ and $K_S^2 = 100 K_X^3$, we get the first part of (2). The second part of (2) follows from the Riemann-Roch formula in Reid [@YPG] and Miyaoka-Yau inequality.
\(3) By the proof of Hacon [@Hac] page 6, under the assumption that $p_g(X) > 0$, one has $\chi(\omega_X) \le 2 p_g(X)$. Now (3) follows from (2).
\(4) Consider first the case $\dim X = 2$. The system $|rK_X|$ is base point free for all $r \ge 4$, by Bombieri [@Bo]. By (1), $\chi({{\mathcal{O}}}_X) \le 1 + p_g(X) \le 3 + K_X^2/2$. The pluri-genus formula says that $P_r(X) = r(r-1)K_X^2/2 + \chi({{\mathcal{O}}}_X) \le 3 +
\frac{1}{2} (1 + r(r-1)) K_X^2$. One can verify that (4) is true.
Now assume that $n \ge 3$. Let $X_{n-1}$ be a general member of $|rK_X|$. By Kollar-Mori [@KM] Lemma 5.17, $X_{n-1}$ is again terminal with $K_{X_{n-1}} = (1 + r) K_X |{X_{n-1}}$. Let $X_i$ be the the intersection of $n-i$ general members of $|rK_X|$. Inductively, we see that each $X_i$ is terminal with nef and big $K_{X_i} = (1 + (n-i)r) K_X|X_{i}$. Also $C = X_1$ is a connected smooth curve; note that $C$ is linearly equivalent to the nef and big divisor $rK_X | X_2$ on the terminal, which is smooth in dimension 2, surface $X_2$; one may also refer to Hartshorne [@Hart] Ch III, Exercise 11.3, for the connectedness of each $X_i$. Inductively, we obtain $h^0(X, rK_X) \le 1 + h^0(X_{n-1},
rK_X |X_{n-1}) \le \dots \le (n-1) + h^0(C, rK_X | C)$, by considering the exact sequence of cohomologies coming from $$0 \rightarrow {{\mathcal{O}}}_{X_{i+1}}
\rightarrow {{\mathcal{O}}}_{X_{i+1}}(rK_X | X_{i+1}) \rightarrow
{{\mathcal{O}}}_{X_i}(rK_X|X_i) \rightarrow 0.$$ The divisor $D = rK_X | C$ on $C$ is special in the sense of Griffiths-Harris [@GH] page 251, because $K_C - D = (1 + (n-2) r) K_X | C \ge (r+1) K_X | C
\ge 0$. Applying Clifford’s Theorem \[ibid\], one has $h^0(C, D) \le
1 + \deg(D)/2$. Substituting this and $\deg(D) = r K_X .
(rK_X)^{n-1} = r^n K_X^n$ into the above, we obtain (4).
In the case of generically finite surjective map between varieties of general type, the volume upstairs can control the volume downstairs and sometimes even the degree of the map. See also Lemma \[Kdeg’\].
\[Kdeg\] Let $X$ and $Y$ be smooth projective $n$-folds of general type and $f : X \rightarrow Y$ a generically finite surjective morphism.
[ ]{}
We have $V(X) \ge \deg(f) \,\, V(Y)$.
Suppose further that $|K_Y|$ defines a generically finite rational map $\Phi$. Then $V(Y) \ge \deg \Phi \,\, \deg \Phi(Y) \ge 1$ and $\deg(f) \le V(X)$.
Let $\varepsilon > 0$. By Fujita’s approximation as in Lazarsfeld [@La] Theorem 11.4.4, after smooth modifications of $X$ and $Y$, we may assume that $K_Y \sim_{\bf Q} H + E$ with $H$ an ample ${\bold Q}$-divisor and $E$ an effective ${\bold Q}$-divisor, such that $H^n = V(H) > V(Y) - \varepsilon$. By the ramification divisor formula, we have $V(X) \ge V(f^*H) = (f^*H)^n = \deg(f)
H^n > \deg(f) (V(Y) - \varepsilon)$. Now (1) follows by letting $\varepsilon$ tend to zero.
The first inequality in (2) is by the proof of Hacon-McKernan [@HM-birat] Lemma 2.2; the rest now follows.
The assertion (1) below means that the volume controls the order of a free-acting group, while (2) says that sometimes $|G|$ can be bounded by its exponent $\exp(G)$. We denote by ${{\text{\rm cartier}}}(Y)$ the Cartier index, i.e., the smallest positive integer such that $c K_Y$ is a Cartier divisor.
\[Kquot\] Let $X$ be a smooth minimal projective $n$-fold of general type and $f : X \rightarrow Y$ a finite surjective morphism.
[ ]{}
Suppose that $f$ is unramified. Then $\deg(f) = K_X^n/K_Y^n \le
K_X^n$. In particular, if a subgroup $H \le {{\text{\rm Aut}}}(X)$ acts freely on $X$, then $|H| = K_X^n / K_Y^n \le K_X^n$, where $Y = X/H$.
Suppose that a subgroup $G \le {{\text{\rm Aut}}}(X)$ has isolated fixed locus $X^G$. Let $f : X \rightarrow Y := X/G$ be the quotient map and $c
= {{\text{\rm cartier}}}(Y)$ the Cartier index. Then $c \, | \, \exp(G)$ and $|G| \le c K_X^n \le \exp(G) \,\, K_X^n$.
We note that $K_X$ is nef and big and that $K_X^n$ is a positive integer.
\(1) follows from the fact that $K_X = f^*K_Y$.
\(2) Denote by $G_x = \{g \in G \, | \, g(x) = x\}$ the stabilizer subgroup at the point $x \in X$. We may regard $G_x$ as a subgroup of ${{\text{\rm GL}}}_n(T_{X, x}) \cong {{\text{\rm GL}}}_n({\bold C})$. For the image $\overline{x} \in Y$ of $x$, the local Cartier index $c({\overline
x})$ equals the index of the quotient group $G_x/G_x \cap
{{\text{\rm SL}}}_n({{\mathbb C}}) \subset {{\text{\rm GL}}}_n({{\mathbb C}})/{{\text{\rm SL}}}_n({{\mathbb C}}) \cong {{\mathbb C}}^*$, which is cyclic and generated by the image $\overline{g}$ of some $g \in
G_x$. Thus $c({\overline x})$ divides $\exp(G_x)$ and also $\exp(G)$. Hence $c = {{\text{\rm cartier}}}(Y)$, which is the ${{\text{\rm lcm}}}$ of $c({\overline x})$, also divides $\exp(G)$. Since $c K_Y$ is Cartier, it can be written as $H_1 - H_2$ with very ample divisors $H_j$ not passing through the isolated set ${{\text{\rm Sing}}}(Y)$ (the image of $X^G$). Then $c K_Y^n = (H_1 - H_2) . K_Y^{n-1} = (K_{Y} |
H_1)^{n-1} - (K_Y | H_2)^{n-1}$ is an integer. Since $f : X
\rightarrow Y$ is etale outside the finite set ${{\text{\rm Sing}}}(Y)$, one has $K_X = f^*K_Y$. Hence $|G| = \deg(f) = K_X^n / K_Y^n = c K_X^n / c
K_Y^n \le c K_X^n$.
An application of the result below is given in Remark \[remBb\]. We remark that the finiteness assumption below on $X^g$ can be removed for surfaces as shown in Ueno [@Ue2].
\[Lefix\] Suppose that $X$ is a smooth projective $n$-fold and $g$ is an automorphism of finite order. Assume that $X^g$ is finite.
[ ]{}
The Euler number of the fixed locus $X^g$ is given by $e(X^g) =$ $\sum_{i=0}^{2n}$ $(-1)^i {{\text{\rm Tr}}}\ g^* | H^i(X, {{\mathbb{Z}}})/({{\text{\rm torsion}}})$.
One has $|e(X^g)| \le \sum_{i=0}^{2n} B_i(X)$, where $B_i(X)$ is the Betti number.
\(1) is the classical Lefschetz fixed point formula. For (2), let ${{\text{\rm ord}}}(g) = m$ and diagonalize $g^* | H^i(X, {{\mathbb C}}) =
{{\text{\rm diag}}}[\zeta_m^{s_1}, \dots, \zeta_m^{s_{b_i}}]$ where $b_i = B_i(X)$ and $\zeta_m^{s_j}$ is an $m$-th root of $1$. Thus $|{{\text{\rm Tr}}}g^* |
H^i(X, {{\mathbb{Z}}})/{{\text{\rm torsion}}}| = |\sum_{j=1}^{b_i} \zeta_m^{s_j}| \le b_i$. This proves the lemma.
The following is the relation between two stabilizers $G_x$ and $(G/H)_{\overline x}$.
\[GGx\] Suppose that a normal subgroup $H \trianglelefteq G$ acts freely on a variety $X$. Set $\overline{G} = G/H$ and $\overline{X} = X/H$.
[ ]{}
For every $x \in X$ and its image $\overline{x} \in \overline{X}$, the homomorphism between stabilizers $\varphi : G_x \rightarrow
G_{\overline x}$ which is given by $g \mapsto \overline{g} = gH$, is an isomorphism.
In particular, the fixed locus $X^G$ is empty (resp. finite) if and only if $\overline{X}^{\overline{G}}$ is empty (resp. finite).
To show that $\varphi$ is surjective, suppose that $\overline{g}
(\overline{x}) = \overline{x}$. Then $g(x) = h(x)$ for some $h \in
H$. Thus $h^{-1} g \in G_x$ and $\overline{g} = \varphi(h^{-1}g)$. To show that ${{\text{\rm Ker}}}(\varphi)$ is trivial, suppose that $\varphi(g)
= {{\text{\rm id}}}\in \overline{G}_{\overline x} \le \overline{G}$. So $g \in
H$. Hence $g \in H \cap G_x = \{{{\text{\rm id}}}\}$ by the freeness of the $H$ action on $X$. So ${{\text{\rm Ker}}}(\varphi)$ is trivial and the lemma is proved.
Next are about nilpotent groups and relation between centralizers.
\[gTh\] Let $G$ be a finite group.
[ ]{}
If $G$ is nilpotent, then $\exp(G)$ equals $\mu(G) := \max$ $\{{{\text{\rm ord}}}(g) \, | \, g \in G\}$.
Consider quotient groups inclusion $K/H \trianglelefteq G/H$. Let $\tau \in K$ and $\overline{\tau} = \tau H \in G/H$. Then $|C_{K/H}({\overline \tau})| \le |C_K(\tau)| \le |C_G(\tau)|$.
\(1) is true for $p$-groups while $G$, being nilpotent, is a direct product of its Sylow subgroups (see Gorenstein [@Go] Theorem 3.5). So (1) follows.
\(2) Write $C_{K/H}({\overline \tau}) = L/H$. Consider the map: $\varphi : L \rightarrow \tau H$ ($g \mapsto g^{-1} \tau g$). Then $|L/C_L(\tau)| = |{{\text{\rm Im}}}(\varphi)| \le |H|$, so $|L/H| \le
|C_L(\tau)|$.
When ${{\text{\rm Aut}}}(X)$ fixes a point $x$, Jordan’s lemma and its consequences below, together with Xiao’s Theorem \[XiaoTh\], will produce a linear bound $|{{\text{\rm Aut}}}(X)| \le J_nx_n K_X^n$.
\[Jordan\] The following are true.
[ ]{}
([**Jordan’s lemma**]{}) Let $J_n = (n+2)!$ (resp. $J_n = n^4 (n+2)!$) for $n > 63$ (resp. for $n \le 63$). Then every finite group in ${{\text{\rm GL}}}_n({{\mathbb C}})$ contains an abelian normal subgroup of index $\le J_n$.
Suppose that $X$ is an $n$-fold and $G \le {{\text{\rm Aut}}}(X)$ is a finite subgroup fixing a smooth point $x \in X$. Then $G$ contains an abelian normal subgroup of index $\le J_n$.
Suppose that $X$ is a projective $n$-fold of general type and $x
\in X$ a smooth point. If $G \le {{\text{\rm Aut}}}(X)$ fixes $x$, then $G$ contains an abelian normal subgroup of index $\le J_n$.
For (1), see Weisfeiler [@We] page 5279 (for better bound) and Aljadeff- Sonn [@AS] page 353. The original number produced by Jordan was $J(n) = (49n)^{n^2}$. For (2), $G$ can be regarded as a subgroup of ${{\text{\rm GL}}}_n(T_{X, x}) \cong {{\text{\rm GL}}}_n({{\mathbb C}})$, so apply (1). For (3), ${{\text{\rm Aut}}}(X)$ and hence $G$ are finite since $X$ is of general type. Then apply (2).
The result below has been proved by many authors. We only mention that the nefness of $\Omega_X^1$ comes from the exact sequence below and the fact that quotient bundles of a nef bundle are again nef (see Hartshorne [@Hart] Ch II, Theorem 8.17; Lazarsfeld [@La] Proposition 6.1.2): $$0 \rightarrow N^{\vee}_{X/A} \rightarrow
\Omega_A^1 | X \rightarrow \Omega_X^1
\rightarrow 0.$$
\[Gauss\] Let $A$ be an abelian variety and $X \subset A$ a subvariety of general type.
[ ]{}
$|K_{\tilde X}|$ defines a generically finite map, where $\tilde{X} \rightarrow X$ is any desingularization.
Suppose that $X$ is smooth. Then $\Omega_X^1$ is nef and $K_X$ is ample. Moreover, $|K_X|$ is base point free and $\Phi_{|K_X|}$ is a finite morphism.
For the proof, see Ran [@Ra] Corollary 2, or Abramovich [@Ab]. See also Hartshorne [@Hart71], Ueno [@Ue] and Griffiths-Harris [@GH79] for earlier development.
Here are some properties of simple abelian varieties to be used in §6.
\[sA\] Let $A$ be a simple abelian variety.
[ ]{}
Every proper subvariety of $A$ is of general type.
If ${{\text{\rm id}}}\ne g \in {{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A)$ is not a translation, then $A^g$ is a non-empty finite set.
If $X \rightarrow A$ is a non-contant morphism from a projective variety $X$, then $q(\tilde{X}) \ge \dim(A)$, where $\tilde{X}
\rightarrow X$ is any desingularization.
\(1) is just Ueno [@Ue] Corollary 10.10.
\(2) Suppose that $g$ is not a translation. By Birkenhake-Lange [@BL] Lemma 13.1.1, we may assume that ${{\text{\rm id}}}\ne g \in
{{\text{\rm Aut}}}_{{{\text{\rm group}}}}(A)$. Now ${{\text{\rm Ker}}}(g - {{\text{\rm id}}})$ is a proper subgroup of $A$, whence it is non-empty and $0$-dimensional by the simplicity of $A$. So (2) follows (see [@BL] Formula 13.1.2).
\(3) By the universal property of ${\text{\rm Alb}}(\tilde{X})$, the non-constant morphism $\tilde{X} \rightarrow A$ factors as $\tilde{X} \rightarrow {\text{\rm Alb}}(\tilde{X}) \rightarrow A$, where the latter map is a homomorphism (modulo a translation) and has image the translation of a non-trivial subtorus of $A$. This image is $A$ by the simplicity of $A$. This implies (3) : $\dim(A) \le
\dim {\text{\rm Alb}}(\tilde{X}) = q(\tilde{X})$.
By verifying the case $\ell = 8$ (resp. $\ell = 12$) and using induction, we have (1) (resp. (2)) below to be used in Section 5.
\[alt\] Let $A_k$ be the alternating group of order $k!/2$. Let $\ell = [k/4] = \max\{z \in {{\mathbb{Z}}}\, | \, z \le k/4\}$ so that $4 \ell \le k \le 4 \ell + 3$.
[ ]{}
If $\ell \ge 8$, then $|A_k| \le (4 \ell +3)!/2 \le |A_{3 \ell}|^{1.7}$.
If $\ell \ge 12$, then $|A_k| \le |A_{\ell}|^8$.
We end §2 with Xiao’s linear bound for abelian subgroups:
\[XiaoTh\] (Xiao [@Xi96] Theorem 1) Let $X$ be a smooth projective $n$-fold with nef and big $K_X$. Let $G \le {{\text{\rm Aut}}}(X)$ be an abelian subgroup. Then there exists a constant $x_n$, independent of $X$, such that $|G| \le x_n K_X^n$.
**G-equivariant fibrations; the proof of Theorem \[Thconj\]**
=============================================================
In this section we consider $G$-equivariant fibration $f : X
\rightarrow Y$. We will prove simultaneously Theorem \[Th3.1\] below and Theorem \[Thconj\] in the Introduction.
\[Th3.1\] Let $X$ and $Y$ be smooth projective varieties of general type and dimensions $n$ and $k$, respectively. Suppose that a group $G$ acts faithfully on $X$ and acts anyhow on $Y$. Let $f : X \rightarrow Y$ be a $G$-equivariant surjective morphism with connected general fibre $F$. Suppose further that $|{{\text{\rm Bir}}}(Y)| \le a_1 \ V(Y)^{d}$ and $|{{\text{\rm Bir}}}(F)| \le a_2 \ V(F)^{d}$ for some positive integers $a_i$ and $d$.
Then we have: $$|G| \le a \, V(X)^d, \hskip 2pc \text{\rm with} \hskip 1pc a = \frac{a_1
a_2}{(\binom{n}{k})^d} \,\, .$$
We now prove first Theorem \[Th3.1\] and late Theorem \[Thconj\] as an application. Note that the general fibre $F$ is also of general type since the Kodaira dimension $\kappa(X) = \dim X$ and by Iitaka’s easy addition: $\kappa(X) \le \kappa(F) + \dim Y$. Since $f$ is $G$-equivariant, we have the natural exact sequence: $$1 \rightarrow K \rightarrow G \rightarrow {{\text{\rm Aut}}}(Y).$$ The group $K$ acts trivially on $Y$ and hence faithfully on the fibre $F$, so $K$ can be regarded as a subgroup of ${{\text{\rm Aut}}}(F)$, where $F$ is smooth and of general type. Now Theorem \[Th3.1\] follows from the Theorem in the Appendix: $$|G| \le |K| \, |{{\text{\rm Aut}}}(Y)| \le |{{\text{\rm Aut}}}(Y)| \ |{{\text{\rm Aut}}}(F)| \le$$ $$|{{\text{\rm Bir}}}(Y)| \ |{{\text{\rm Bir}}}(F)|
\le a_1 a_2 \ V(Y)^d \ V(F)^d \le a V(X)^d.$$
Finally, Theorem \[Thconj\] is a consequence of Theorem \[Th3.1\] and the inductive assumption, where we set $a_1 =
\delta_{\dim Y}$, $a_2 = \delta_{\dim F}$ and $d = 1$.
Weaker versions of Theorems \[Thconj\] and \[Th3.1\] can also be proved using Lemma \[smalldiminv\], Kollar [@KoI] Theorem 3.5 (ii) and Catanese-Schneider [@CS] pages 10-11, which are strengthened to the current form, thanks to the Appendix - an application of the weak positivity due to Fujita [@Fuj], Kawamata [@Ka82b] and Viehweg [@Vi82b].
**Bound of Betti numbers; the proof of Theorem \[Bbetti\]**
===========================================================
In this section, for a smooth projective variety $X$, we shall bound the invariants like Betti numbers $B_i(X)$ and the Euler number $e(X)$, in terms of the volume of $X$. We first bound the irregularity and geometric genus.
\[qp\] Let $X$ be a smooth projective $n$-fold ($n \ge 2$) with ample $K_X$. Then $p_g(X) \le e_n K_X^n$ and $q(X) \le e_n K_X^n$, where $e_n = \max\{n + r^n, \, 2 + \frac{1}{2} r^{n-2} (1 + (n-2) r)^2\}$ with $r = 2 + n(n+1)/2$.
For $r = 2 + n(n+1)/2$, as in Lemma \[smalldiminv\], $|rK_X|$ is base point free. Since $K_X$ is ample, $\Phi = \Phi_{|rK_X|} : X
\rightarrow {{\mathbb P}}^N$ with $N = P_r(X) - 1$, is a finite morphism. Hence $(rK_X)^n = \deg(\Phi) \deg \Phi(X) \ge \deg \Phi(X) \ge N -
n + 1$ (see Griffiths-Harris [@GH] page 173). Thus $p_g(X) \le
P_r(X) = N + 1 \le n + r^n K_X^n \le (n + r^n) K_X^n$.
As in Lemma \[smalldiminv\], take the ladder $X = X_{n} \supset X_{n-1} \supset \cdots \supset X_2$, where $X_i$ is the intersection of $n-i$ general members of $|rK_X|$, and a smooth $i$-fold with ample $K_{X_i} = (1 + (n-i)r) K_X | X_i$. Each $X_{i-1}$ is a smooth ample divisor on $X_i$. By Lefschetz hyperplane section theorem, we have $q(X) = q(X_{n-1}) = \dots = q(X_2)$. By Lemma \[smalldiminv\], we conclude the proposition: $q(X_2) \le \frac{1}{2}K_{X_2}^2 + 2
= 2 + \frac{1}{2} r^{n-2} (1 + (n-2) r)^2 K_X^n$.
To bound the Betti numbers, we need the bound of Euler number $e(X)$ below.
\[eX\] Let $X$ be a smooth projective $n$-fold.
[ ]{}
If the cotangent bundle $\Omega_X^1$ is nef, then $0 \le (-1)^n e(X) = c_n(\Omega_X^1) \le K_X^n$.
If $K_X$ is ample then $|e(X)| \le h_n K_X^n$, where $h_n$ is the constant, independent of $X$, in Heier [@He] Proposition 1.7.
\(2) is proved in Heier [@He] Proposition 1.7. For (1), by the nefness of $\Omega_X^1$ and Demailly-Peternell-Schneider [@DPS] page 31, we have $0 \le c_n(\Omega_X^1) \le c_1(\Omega_X^1)^n =
K_X^n$. Note also that $c_n(\Omega_X^1) = (-1)^n c_n(T_X) = (-1)^n
c_n(X) = (-1)^n e(X)$, where $T_X$ is the tangent bundle; see for instance, Fulton [@Ful] Example 3.2.13. This proves the lemma.
[**Proof of Theorem \[Bbetti\]**]{}
We will prove by induction on $n$. The case $n = 1$ is clear. When $n = 2$, it follows from Lemma \[smalldiminv\] or \[qp\]. Indeed, note that $B_1(X) = 2q(X)$, $\chi({{\mathcal{O}}}_X) \le 1 + p_g(X)$, $c_2(X) \le K_X^2 + c_2(X) = 12 \chi({{\mathcal{O}}}_X)$ and $B_2(X) = c_2(X)
- 2 + 4q(X)$.
Suppose $n \ge 3$ and the theorem is true for all such $k$-folds with $2 \le k < n$. Let $X_{n-1} $ be a general member of $|rK_X|$ with $r = 2 + n(n+1)/2$ (see Lemma \[smalldiminv\]). Note that $K_{X_{n-1}} = (1 + r)K_X | X_{n-1}$ is ample and $K_{X_{n-1}}^{n-1} = r(1 + r)^{n-1} K_X^n$. By induction, $B_i(X_{n-1}) \le a_{n-1} K_{X_{n-1}}^{n-1} = a_n' K_X^n$ where $a_n' = a_{n-1} r (1+r)^{n-1}$. By Lefschetz hyperplane section theorem, $B_i(X) = B_i(X_{n-1})$ for all $i \le n-2$ and $B_{n-1}(X) \le B_{n-1}(X_{n-1})$ (see Lazarsfeld [@La] Theorem 3.1.17). Note also that $B_j(X) = B_{2n-j}(X)$ by Lefschetz duality. Thus $B_{j}(X) \le B_{j}(X_{n-1}) \le a_n'
K_X^n$ for all $j \ne n$.
On the other hand, by Lemma \[eX\], $h_n K_X^n \ge
|\sum_{i=0}^{2n} (-1)^{i} B_i(X)|$. So $|B_n(X)| \le h_n K_X^n + |\sum_{i \ne n} (-1)^{i} B_i(X)|
\le a_n K_X^n$ where $a_n = h_n + 2n a_n'$. We are done by the induction. This proves Theorem \[Bbetti\].
\[remBb\] For the $X$ in Theorem \[Bbetti\], it might be possible to give a second proof of Xiao’s linear bound (in terms of $K_X^n$) of ${{\text{\rm ord}}}(g)$ for every $g \in {{\text{\rm Aut}}}(X)$ with ${{\text{\rm ord}}}(g)$ prime and $X^g$ finite. Indeed, note that the quotient map $X' = X - X^g \rightarrow
X'/\langle g \rangle$ is unramified, and hence ${{\text{\rm ord}}}(g)$ divides the Euler number $e = e(X) - e(X^{g})$, so $e$ is bounded by $|e(X)| +
|e(X^g)|$ (provided that $e \ne 0$) while the latter is bounded linearly by $K_X^n$ (see Lemmas \[Lefix\] and \[eX\] and Theorem \[Bbetti\]).
**Automorphism groups of subvarieties of abelian varieties**
============================================================
In this section, we shall prove the following two results:
\[Th2’\] Let $X$ be a smooth $n$-fold ($n \ge 3$) of general type contained in an abelian variety $A$ of dimension $q$. Let $G$ be a subgroup of $\{g \in {{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A) \, | \, g(X)$ $= X\}$ such that the fixed locus $A^g$ for every ${{\text{\rm id}}}\ne g \in G$ is a non-empty finite set unless $g$ is a translation of $A$. Set $V = K_X^n$ (see Theorem \[Gauss\]).
Then there is a constant $d_n$, independent of $X$ and $A$, such that $|G| \le d_n q^{10-b}V^b \le d_n S^{10}$, where $S = \max\{V,
q\}$ and $5 \le b \le 10$.
\[Th2”\] Let $A$ be a simple abelian variety and $X \subset A$ a smooth $n$-dimensional ($n \ge 3$) proper subvariety. Let $G = \{g \in {{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A) \, | \,$ $g(X)= X\}$ be the stabilizer. Set $V = K_X^n$ (see Theorem \[Gauss\]). Then there is a constant $d_n'$ (independent of $X$ and $A$) such that $|G| \le d_n' V^{10}$.
We fix an abelian variety $A$ of dimension $q$ and an $n$-dimensional smooth subvariety $X \subset A$ of general type. Let $G$ be a subgroup of the group $\{g \in {{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A) \, | \,
g(X) = X\}$. Note that ${{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A) = T \rtimes A_0$ (split extension) where $T = \{T_t \, | \, t \in A\}$ is the normal subgroup of translations and $A_0 = {{\text{\rm Aut}}}_{{{\text{\rm group}}}}(A)$ is the subgroup of bijective group-homomorphisms.
\[setup1’\] [**Assumption**]{}.
[ ]{}
$X \subset A$ is a smooth projective $n$-fold of general type in the abelian variety $A$ of dimension $q$,
the subgroup $G \subseteq \{g \in {{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A) \, | \, g(X) = X\}$ contains no translations of $A$ (we note that $G \le {{\text{\rm Aut}}}(X)$ is finite because $X$ is of general type), and
the fixed locus $A^g$ for every ${{\text{\rm id}}}\ne g \in G$ is a non-empty finite set.
We collect some information on the structure of $G$.
\[gr\] Suppose that $G \subseteq {{\text{\rm Aut}}}(X)$ and $X \subset A$ satisfy conditions in \[setup1’\]. Let ${{\text{\rm id}}}\ne g \in G$. Then we have:
[ ]{}
Let $G_0$ be the image of the composition $G \subseteq {{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A)
\rightarrow A_0$. Then $G \rightarrow G_0$ is an isomorphism.
For $g \in G$ and $g_0$ its image in $G_0$, we have $|A^g| = |A^{g_0}| \ge |\{0\}| = 1$.
If $H \le G$ is abelian, then $H$ is cyclic.
For $H \le G$, if $|H| = p^2$ for some prime $p$, then $H$ is cyclic.
If $|X^g| \ge 2$ for some $g \in G$, then ${{\text{\rm ord}}}(g) = p^s$ for some prime $p$.
If $g$ has order $m$ in $G$ and if $\zeta_m$ denotes a primitive $m$-th root of $1$, then there is is a diagonalization $g^* | H^0(A,
\Omega_A^1)$ $= {{\text{\rm diag}}}[\zeta_m^{s_1}, \dots, \zeta_m^{s_q}]$ where each $\zeta_m^{s_j}$ is a primitive $m$-th root of $1$.
If $g \in G$, then the Euler function $\varphi({{\text{\rm ord}}}(g))$ divides $2q$.
If $p^2$ divides $|G|$ for some prime $p$, then there is an element $g \in G$ such that ${{\text{\rm ord}}}(g) = p^2$.
If $g \in G$, then ${{\text{\rm ord}}}(g)$ divides $4q^2 \alpha(g)$, where $$\alpha(g) = \prod_{\text{\rm prime}
\, p \,|\, {{\text{\rm ord}}}(g), \, p^2 \nmid |G|} \, p.$$
If $H \le G$, then the exponent $\exp(H)$ divides $4q^2
\alpha(H)$, where $$\alpha(H) = \prod_{\text{\rm prime}
\, p \,|\, |H|, \, p^2 \nmid |G|} \, p.$$
\[setup1’\] (ii) implies (1). Here (2), (3), (6) and (7) are consequence of \[setup1’\] (iii) and Birkenhake - Lange [@BL] Lemma 13.1.1 and Propositions 13.2.2, 13.2.1 and 13.2.5. Our (4) is because a group of order $p^2$ is abelian and by (3). Now (5) follows from (1) and Birkenhake - Lange [@BL] Corollary 13.2.4. Also (8) is a consequence of (4) together with Sylow’s theorem and a basic result on $p$-groups. For (9), write ${{\text{\rm ord}}}(g) = p_1^{t_1}
p_2^{t_2} \dots p_u^{t_u}$ with $t_i \ge 1$. Then $\varphi({{\text{\rm ord}}}(g))
= p_1^{t_1-1}(p_1 - 1) \dots p_u^{t_u - 1}(p_u - 1)$. If $t_i \ge
2$, then $2(t_i - 1) \ge t_i$, and $p_i^{t_i}$ divides $(\varphi({{\text{\rm ord}}}(g)))^2$ as well as $(2q)^2$ by (7). If $p^2 \, | \,
|G|$, then $p \, | \, 2q$ by (8) and (7). Thus ${{\text{\rm ord}}}(g) \, | \, 4q^2
\alpha(g)$ and (9) is proved. Our (9) implies (10). This proves the lemma.
The centralizer lemma below is a key to the proof of Theorem \[Th2’\].
\[central\] ([**centralizer lemma**]{}) Suppose that $G \subseteq {{\text{\rm Aut}}}(X)$ and $X \subset A$ satisfy conditions in \[setup1’\]. Let $\tau \in G$ be of prime order $p$ such that the fixed locus $X^{\tau}$ is non-empty. Let $V = K_X^n$ (see Theorem \[Gauss\]).
[ ]{}
Every prime factor $p_1 \ne p$ of the order $|C_G(\tau)|$ of the centralizer, divides $|X^{\tau}| (|X^{\tau}| - 1)$.
$\exp(C_G(\tau))$ divides $4 p^{\varepsilon} q^2|X^{\tau}|
(|X^{\tau}| - 1)$, where $\varepsilon = 0$ (resp. $1$) when $p^2$ divides $|G|$ (resp. otherwise).
One has $|C_G(\tau)| \le k_n q^2V^{3 + \varepsilon}$, where $k_n =
\max\{J_n x_n, \, 4a_n^2 x_n^{\varepsilon}\}$; see Lemma \[Jordan\] and Theorems \[Bbetti\] and \[XiaoTh\].
\(1) Suppose $g \in C_G(\tau)$ has order equal to a prime $p_1 \ne p$. Write $|X^{\tau}| \equiv r$ with $0 \le r < p_1$. If $r = 0, 1$, then we are done. Suppose that $r \ge 2$. Then $|X^{\tau g}| \ge r \ge 2$. By Lemma \[gr\], $p p_1 = {{\text{\rm ord}}}(\tau g)$ equals a prime power, a contradiction.
\(2) follows from (1) and Lemma \[gr\] (10).
\(3) If $|X^{\tau}| = 1$, then $C_G(\tau)$ fixes the unique point $x$ in $X^{\tau}$. Hence $C_G(\tau) \le {{\text{\rm GL}}}_n(T_{X, x})$ and we apply Lemma \[Jordan\] and Theorem \[XiaoTh\] and obtain $|C_G(\tau)| \le J_n x_n K_X^n$. Suppose now that $|X^{\tau}| \ge
2$. By (2), Lemmas \[Kquot\] and \[Lefix\] and Theorem \[Bbetti\], we have $|C_G(\tau)| \le \exp(C_G(\tau) V \le 4
p^{\varepsilon} q^2|X^{\tau}| (|X^{\tau}| - 1) V \le 4
p^{\varepsilon} q^2 (\sum_{i=0}^{2n} B_i(X))^2 V \le 4a_n^2
p^{\varepsilon} q^2 V^3$. By Theorem \[XiaoTh\], $p \le x_n V$ and hence (3) follows. This proves the lemma.
Theorem \[Th2’\] should follow from the crucial proposition below.
\[key\] Suppose that $X \subset A$ and $G \subseteq {{\text{\rm Aut}}}(X)$ satisfy the conditions in \[setup1’\]. Let $q = \dim(A)$ and $V = K_X^n$ (see Theorem \[Gauss\]). Then the there is a constant $d_n$, independent of $X$ and $A$, such that $|G| \le d_n q^{10-b} V^b \le
d_n S^{10}$, where $S = \max\{V, q\}$ and $5 \le b \le 10$.
For the proof, we argue closely along the lines of Huckleberry-Sauer [@HS] and Szabo [@Sz] (see [@As]). But we use the centralizer Lemma \[central\] instead. Let $H
\trianglelefteq G$ be a maximum among normal subgroups of $G$ which acts freely on $X$. Take a normal subgroup $K/H
\trianglelefteq G/H$. Fix some $\tau \in K$ such that the fixed locus $X^{\tau} \ne \emptyset$. We may assume that ${{\text{\rm ord}}}(\tau) =
{{\text{\rm ord}}}(\overline{\tau}) = p$ is a prime (see \[setup1’\] (iii) and Lemma \[GGx\]).
\[GG1\] Let $G_1 = C_G(\tau)$. Then $|G| \le |G_1| |K|$.
Note that $\varphi : G \rightarrow K$ ($g \mapsto g^{-1} \tau g$) is a well-defined map. Then the claim follows from that $|G/G_1| = |\varphi(G)| \le |K|$.
Suppose that we can choose $K/H$ to be abelian. Note that $K/H$ acts on the smooth minimal $n$-fold $X/H$ of general type. By Theorem \[XiaoTh\], we have $|K/H| \le x_n K_{X/H}^n
= x_n V/|H|$. Thus by Claim \[GG1\] and Lemmas \[central\] and \[Kquot\], we have $|G| \le |G_1| |K/H| |H| \le k_nq^2 V^{3 + \varepsilon} x_n V
\le d_n q^2 V^{5}$ with $d_n = k_n x_n$.
We may assume that there is no such abelian $K/H$. Then $G/H$ is semi-simple, i.e., it has no non-trivial abelian normal subgroup. Let $M/H$ be a non-trivial minimal normal subgroup. Then $M/H = \prod^k E/H$, a $k$-fold product of the same non-abelian simple group $E/H$. Let $M_i/H$ be all distinct non-trivial minimal normal subgroups of $G/H$ and let $S/H = \prod M_i/H$ be the sockel of $G/H$. We write $M_i/H = \prod^{k_i} E_i/H$ with non-abelian simple group $E_i/H$.
Suppose that $M/H = \prod^k E/H$ with $k \ge 2$. Then $\alpha(M)
\, | \, |H|$ in notation of Lemma \[gr\], because for every prime factor $p_1$ of $|M|$, either $p_1 \, | \, |H|$, or $p_1 \,
| \, |M/H|$ and hence $p_1^k \, | \, |M/H|$ (and $p_1^2 \, | \,
|G|$). So by Lemma \[gr\], $\exp(M)$ divides $4q^2 \alpha(M)$ and also $4q^2 |H|$. Our Lemma \[Kquot\] implies that $|M| \le
\exp(M) K_X^n$. Substituting all these in and applying Lemmas \[central\] and \[Kquot\], we obtain $|G| \le |G_1| |M| \le
k_nq^2 V^{3 + \varepsilon} 4q^2 |H| V \le 4k_n q^4 V^6$ and we are done.
So we may assume that $M/H = E/H$ is non-abelian simple for every non-trivial minimal normal $M/H \trianglelefteq G/H$. Suppose that $M/H$ is one of the $26$ sporadic non-abelian simple groups. Then $M/H \le d$, a constant (the order of the Monster simple group). Thus as above, $|G| \le |G_1| |M/H| |H| \le k_nq^2 V^{3 + \varepsilon}
d V \le d k_n q^2 V^5$.
Suppose that $E/H$ is of Lie type. By the proof of Huckleberry-Sauer [@HS] Proposition 7, there exist a universal constant $d$ and a Sylow $p_1$-subgroup $U/H$ of $M/H$ such that $|M/H| \le d |U/H|^{5/2}$. Note that $U/H$ acts on the smooth minimal $n$-fold $X/H$ of general type and the fixed locus $(X/H)^{U/H}$ is finite by \[setup1’\] (iii) and Lemma \[GGx\]. By Lemmas \[Kquot\] and \[gTh\] and Theorem \[XiaoTh\], we have $|U/H| \le \exp(U/H) K_{X/H}^n = \mu(U/H)
V/|H| \le x_n K_{X/H}^n V/|H| = x_n (V/|H|)^2$. Thus $|M/H| \le d
x_n^{5/2}$ $(V/|H|)^5$. By Lemma \[central\] we have $|G| \le
|G_1| |M| \le k_nq^2 V^{3 + \varepsilon} d x_n^{5/2} V^5$ $= d_n
q^2 V^{8 + \varepsilon}$ with $d_n = k_n d x_n^{5/2}$. If $p^2 \,
| \, |G|$ then $\varepsilon = 0$ and we are done.
If the fixed locus $X^U = \emptyset$, then $|U| \le V$ by Lemma \[Kquot\] and we will even have a better bound. We may assume that the fixed locus $X^U \ne \emptyset$ (equivalently $(X/H)^{U/H} \ne \emptyset$ by Lemma \[GGx\]). Then our initial $\overline{\tau}$ (of order $p$) can be taken from $U/H$ so that $p_1 = p$ and $U/H$ is a $p$-Sylow subgroup of $M/H$. If $p^2$ divides $|U/H|$ then it also divides $|G|$, whence $\varepsilon = 0$ and we are done. Otherwise, $|U/H| = p \le x_n K_{X/H}^n = x_n V/|H|$ by Theorem \[XiaoTh\] and we will be done again.
We are left with the case where each $M_i/H$ is an alternating group $A_{k_i}$. Suppose that $M/H = A_k$ is the smallest among them. If there are two $M_i/H$ then we will be done as in the case $M/H = \prod^k E/H$ with $k \ge 2$ because the fact that $|M/H|$ divides $|M_i/H|$ for two $i$ implies that $\alpha(M) \, | \, |H|$ as well.
Therefore, we assume that the sockel $S/H = M/H = A_k$. Then as noticed by Huckleberry and Sauer, the conjugation map induces an injection $1 \rightarrow G/H \rightarrow {{\text{\rm Aut}}}(S/H)$. Hence we have the following, where the first factor 2 is needed only when $k = 6$ (so that ${{\text{\rm Aut}}}(A_6)/A_6 = ({{\mathbb{Z}}}/(2))^{\oplus 2}$; see Atlas [@Atlas]) : $$|G/H| \le |{{\text{\rm Aut}}}(S/H)| \le 2 \times 2 \, |A_k| = 4 |M/H|.$$ First, treat the case $k \le 58$. Then $|G| \le 4|A_k| |H| \le
4|A_k| V \le 4(58!)V$ by Lemma \[Kquot\], and we are done. Or as noticed by Szabo [@Sz], $M/H$ contains a Sylow $p_1$-subgroup $U/H$ such that (as above) $|M/H| \le |U/H|^5 \le (x_n
(V/|H|)^2)^5$, whence $|G| \le d_nV^{10}$ with $d_n = 4 x_n^5$.
We next deal with the case $k \ge 59$. We use the approach of Szabo [@Sz], but we apply the centralizer Lemma \[gTh\] instead. As in Lemma \[alt\], set $\ell = [k/4]$ and we have $A_{\ell} \times A_{3\ell} \le A_k \le A_{4\ell + 3}$.
Suppose that the subgroup $A_{\ell} < A_k = M/H$ acts freely on $X/H$. Then $\tilde{A}_{\ell} < G$ (with $\tilde{A}_{\ell}/H =
A_{\ell}$) also acts freely on $X$ by Lemma \[GGx\], whence $|\tilde{A}_{\ell}| \le K_X^n = V$ by Lemma \[Kquot\]. By Lemma \[alt\], $|A_k| \le |A_{\ell}|^8 \le (V/|H|)^8$, so $|G| = |G/H|
|H| \le 4 |A_k| |H| \le 4V^8$.
Suppose that $A_{\ell}$ does not act freely on $X/H$. We may take $\overline{\tau} \in G/H$ to be from $A_{\ell}$ and $\tau \in M$ with ${{\text{\rm ord}}}(\tau) = p = {{\text{\rm ord}}}(\overline{\tau})$. Now $A_{3\ell} \le C_{M/H}(\overline{\tau}) \le C_{G/H}(\overline{\tau})$. This and Lemma \[gTh\] imply $|A_{3\ell}| \le |C_G(\tau)| = |G_1|$. By Lemmas \[alt\], \[Kquot\] and \[central\], $|G| \le 4|A_k| |H| \le 4 |A_{3 \ell}|^{1.7} V
\le 4 |G_1|^{1.7} V \le 4(k_n q^2V^{3 + \varepsilon})^{1.7} V
= d_n q^{3.4} V^{6.1 + 1.7 \varepsilon}$, where $d_n = 4 k_n^{1.7}$. Since $p$ divides $|A_{\ell}|$, our $p^2$ divides $A_k = M/H$ and also $|G|$. So $\varepsilon = 0$ and we are done. This completes the proof of Proposition \[key\].
[**Proof of Theorem \[Th2’\]**]{}
Since $X$ is of general type, $G \le {{\text{\rm Aut}}}(X)$ is finite. Write ${{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A) = T \rtimes A_0$ as at the beginning of the section. Set $T_G = T \cap G$ which acts freely on $A$ and $X$. If $G = T_G$, then $|G| \le K_X^n$ by Lemma \[Kquot\], and we are done. So assume that $T_G < G$. Note that $A \rightarrow A/T_G$ is an isogeny of abelian varieties, $X \rightarrow X/T_G$ is etale and $G/T_G \le \{g \in {{\text{\rm Aut}}}_{{{\text{\rm variety}}}}(A/T_G) \, | \, g(X/T_G) =
X/T_G\}$. We shall check the conditions in \[setup1’\] for $G/T_G
\le {{\text{\rm Aut}}}(X/T_G)$ and $X/T_G \subset A/T_G$. For every $g \in G
\setminus T_G$, we have $A^g \ne \emptyset$ by the assumpton on $G$, so $(A/T_G)^{\overline g} \ne \emptyset$ by Lemma \[GGx\]; here $\overline{g} = g T_G \in G/T_G$. Thus $\overline{g}$ is not a translation on $A/T_G$. So all conditions in \[setup1’\] are satisfied by $G/T_G$ and $X/T_G \subset A/T_G$; see Lemma \[GGx\] and the assumption on $G$. By Proposition \[key\], we have $|G/T_G| \le d_n q(A/T_G)^{10-b} (K_{X/T_G}^n)^b = d_n q^{10-b}
(V/|T_G|)^b$. Now the theorem follows because $b \ge 5$. This proves Theorem \[Th2’\].
[**Proof of Corollary \[Th2”\]**]{}
By Lemma \[sA\], $X$ is of general type and the fixed locus $A^g$ for every ${{\text{\rm id}}}\ne g \in G$ is a non-empty finite set unless $g$ is a translation of $A$. Thus we can apply Theorem \[Th2’\]. Now Corollary \[Th2”\] with $d_n' = d_n e_n^5$, follows from Theorem \[Th2’\], Lemma \[sA\], Proposition \[qp\] and Theorem \[Gauss\].
**Proofs of Theorems \[Th1\] and \[Th2\]**
==========================================
In this section we shall prove Theorems \[Th1\] and \[Th2\]. We start with an observation which bounds $\deg({\text{\rm alb}}_X)$ in terms of the volume of $X$. Note that ${\text{\rm Alb}}(X) = {\text{\rm Alb}}(X')$ if $X$ and $X'$ are smooth and birational. The lemma below follows from Lemma \[Kdeg\] and Theorem \[Gauss\] (1) after resolving singularities of $W$, since the volumes are birational invariants.
\[Kdeg’\] Let $X$ be a smooth projective $n$-fold. Suppose that ${\text{\rm alb}}_X : X
\rightarrow W := {\text{\rm alb}}(X) \subset {\text{\rm Alb}}(X)$ is generically finite onto $W$, and $W$ is of general type. Then $V(X) \ge
\deg({\text{\rm alb}}_X) \,\, V(W) \ge \deg({\text{\rm alb}}_X)$.
[**Proof of Theorem \[Th2\]**]{}
Let $X$ and $G = {{\text{\rm Bir}}}(X)$ be as in Theorem \[Th2\]. By Hanamura [@Han] Lemma 2.4, after a smooth modification of $X$, we may assume that $G$ acts regularly on $X$. The universal property of ${\text{\rm Alb}}(X)$ implies that there is an induced action (not necessarily faithful) of $G$ on ${\text{\rm Alb}}(X)$ such that ${\text{\rm alb}}_X : X \rightarrow
{\text{\rm Alb}}(X)$ is $G$-equivariant. Let $K$ (resp. $\overline{G}$) be the kernel (resp. image) of the homomorphism $G = {{\text{\rm Aut}}}(X) \rightarrow
{{\text{\rm Aut}}}_{{{\text{\rm variety}}}}({\text{\rm Alb}}(X))$ so that we have the exact sequence $$1 \rightarrow K \rightarrow G
\rightarrow \overline{G} \rightarrow 1.$$ Since $K$ acts trivially on $W = {\text{\rm alb}}_X(X)$, we can factor ${\text{\rm alb}}_X$ as $X \rightarrow X/K
\rightarrow W$. In particular, $|K| \le \deg({\text{\rm alb}}_X)$. By Lemma \[sA\] and the assumption in Theorem \[Th2\], the fixed locus ${\text{\rm Alb}}(X)^{\overline g}$ for every ${{\text{\rm id}}}\ne \overline{g} \in
\overline{G}$ is a non-empty finite set unless $\overline{g}$ is a translation of ${\text{\rm Alb}}(X)$. Thus, by Theorem \[Gauss\] (2), we can apply Proposition \[qp\] and Theorem \[Th2’\] to $W$. Setting $V
= V(X)$ and $V(W) = K_W^n$ and noting that $q: = q(W) = q(X) = \dim
{\text{\rm Alb}}(X)$ by the universal property and definition of ${\text{\rm Alb}}(X)$, one obtains the inequalities below with $d_n'' = d_n e_n^{10-b}$: $$|\overline{G}| \le d_n q^{10-b} V(W)^b \le d_n'' V(W)^{10}.$$ Now by Lemma \[Kdeg’\] or \[Kdeg\], we conclude Theorem \[Th2\]: $$|G| = |K| |\overline{G}| \le \deg({\text{\rm alb}}_X) |\overline{G}|
\le d_n'' V \cdot V(W)^{9} \le d_n'' V^{10}.$$
[**Proof of Theorem \[Th1\]**]{}
Theorem \[Th1\] (1) is a special case of Theorem \[Th1\] (2). Therefore, we have only to show Theorem \[Th1\] (2).
So suppose that ${\text{\rm alb}}_X$ is not generically finite onto a 3-fold of general type. Set $W = {\text{\rm alb}}_X(X) \subset {\text{\rm Alb}}(X)$. Then either $\dim
W < 3$, or $\kappa(W) < \dim W \le 3$. As above, after a smooth modification of $X$, we may assume that $G = {{\text{\rm Bir}}}(X)$ acts regularly on $X$ and of course on ${\text{\rm Alb}}(X)$ (not necessarily faithful on the latter) so that ${\text{\rm alb}}_X : X \rightarrow {\text{\rm Alb}}(X)$ is $G$-equivariant. Since $\dim {\text{\rm Alb}}(X) = q(X) \ge 4 > \dim X \ge \dim W$ by the assumption, our $W$ is a proper subvariety of ${\text{\rm Alb}}(X)$. Hence by Ueno [@Ue] Lemma 9.14 and Corollary 10.4, the Kodaira dimension $\kappa(W) \ge 1$. As in Ueno [@Ue] Theorem 10.9 or Mori [@Mo87] Theorem 3.7, let $B$ be the identity connected component of $\{a \in {\text{\rm Alb}}(X) \, | \, a + W \subseteq W\}$. Then $W/B$ is of general type, $W \rightarrow W/B$ is an etale fibre bundle with fibre $B$ and is birational to the Iitaka fibring of $W$. Since the pluri-canonical systems of $W$ are $G$-stable, we may take $G$-equivariant desingularizations $X' \rightarrow X$ and $Y'
\rightarrow W/B$ such that $X' \rightarrow Y'$ is a well-defined $G$-equivariant morphism, though $G$ might not act faithfully on $Y'$. Rewrite $X'$ as $X$. Take a Stein factorization $X \rightarrow
Y \rightarrow Y'$ so that $f: X \rightarrow Y$ has connected general fibre $F$ and is necessarily $G$-equivariant. We may also assume that $Y$ is already smooth (or do equivariant modifications again). Note that $G = {{\text{\rm Bir}}}(X) = {{\text{\rm Aut}}}(X)$ now. We have $\dim Y = \dim Y' =
\kappa(W) = 1, 2$. Since the subvariety $W/B < {\text{\rm Alb}}(X)/B$ is of general type, both $Y$ and $Y'$ are of general type. Apply Theorem \[Th3.1\] to the $G$-equivariant map $f : X \rightarrow Y$ with connected general fibre $F$ say. Note that $3 = \dim X = \dim Y +
\dim F$ and $|{{\text{\rm Bir}}}(Z)| = |{{\text{\rm Aut}}}(Z_{\min})| \le (42)^{\dim Z} V(Z)$ for both $Z = Y$ and $Z = F$, thanks to Hurwitz and Xiao. Here $Z_{\min}$ is the [*unique*]{} smooth minimal model of $Z$, since $\dim Z \le 2$. By Theorem \[Th3.1\], we have $|G| \le a \ V(X)$ with $a = (42)^3/3$. This proves Theorem \[Th1\] (2).
[99]{} D. Abramovich, Subvarieties of semiabelian varieties. Compositio Math. 90 (1994), no. 1, 37–52. U. Angehrn and Y. Siu, Effective freeness and point separation for adjoint bundles. Invent. Math. 122 (1995), no. 2, 291–308. E. Aljadeff and J. Sonn, Bounds on orders of finite subgroups of ${\rm PGL}\sb n(K)$. J. Algebra 210 (1998), no. 1, 352–360. M. Aschbacher, The status of the classification of the finite simple groups. Notices Amer. Math. Soc. 51 (2004), no. 7, 736–740. W. Barth, C. Peters and A. Van de Ven, Compact complex surfaces. Ergebnisse der Mathematik und ihrer Grenzgebiete (3), 4. Springer-Verlag, Berlin, 1984. C. Birkenhake and H. Lange, Complex abelian varieties. Second edition. Grundlehren der Mathematischen Wissenschaften, 302. Springer-Verlag, Berlin, 2004. E. Bombieri, Canonical models of surfaces of general type. Inst. Hautes Études Sci. Publ. Math. No. 42, (1973), 171–219. J. Cai, On abelian automorphism groups of $3$-folds of general type. Algebra Colloq. 2 (1995), no. 4, 373–382. F. Catanese and M. Schneider, Polynomial bounds for abelian groups of automorphisms. Special issue in honour of Frans Oort. Compositio Math. 97 (1995), no. 1-2, 1–15. Jungkai A. Chen and C. D. Hacon, Linear series of irregular varieties. Algebraic geometry in East Asia (Kyoto, 2001), 143–153, World Sci. Publishing, River Edge, NJ, 2002. M. Chen, Inequalities of Noether type for 3-folds of general type. J. Math. Soc. Japan 56 (2004), no. 4, 1131–1155. J. H. Conway, R. T. Curtis, S. P. Norton, R . A. Parker and R. A. Wilson, Atlas of finite groups. Oxford University Press, Eynsham, 1985. A. Corti, Polynomial bounds for the number of automorphisms of a surface of general type. Ann. Sci. École Norm. Sup. (4) 24 (1991), no. 1, 113–137. J. -P. Demailly, T. Peternell and M. Schneider, Compact complex manifolds with numerically effective tangent bundles. J. Algebraic Geom. 3 (1994), no. 2, 295–345. S. Friedland, The maximal orders of finite subgroups in ${\rm GL}\sb n(Q)$. Proc. Amer. Math. Soc. 125 (1997), no. 12, 3519–3526. T. Fujita, On Kahler fiber spaces over curves. J. Math. Soc. Japan 30 (1978), no. 4, 779–794. W. Fulton, Intersection theory. Ergebnisse der Mathematik und ihrer Grenzgebiete (3), 2. Springer-Verlag, Berlin, 1984. D. Gorenstein, Finite groups. Harper $\&$ Row, Publishers, New York-London 1968. P. Griffiths and J. Harris, Principles of algebraic geometry. John Wiley $\&$ Sons, Inc., New York, 1994. P. Griffiths and J. Harris, Algebraic geometry and local differential geometry. Ann. Sci. École Norm. Sup. (4) 12 (1979), no. 3, 355–452. C. D. Hacon, On the degree of the canonical maps of 3-folds. Proc. Japan Acad. Ser. A Math. Sci. 80 (2004), no. 8, 166–167. C. D. Hacon and J. McKernan, Boundedness of pluricanonical maps of varieties of general type, math.AG/0504327. M. Hanamura, The birational automorphism groups and the Albanese maps of varieties with Kodaira dimension zero. J. Reine Angew. Math. 411 (1990), 124–136. R. Hartshorne, Ample vector bundles on curves. Nagoya Math. J. 43 (1971), 73–89. R. Hartshorne, Algebraic geometry. Graduate Texts in Mathematics, No. 52. Springer-Verlag, New York-Heidelberg, 1977. G. Heier, Effective finiteness theorems for maps between canonically polarized compact complex manifolds. Math. Nachr. 278 (2005), no. 1-2, 133–140. A. Howard and A. J. Sommese, On the orders of the automorphism groups of certain projective manifolds. Manifolds and Lie groups (Notre Dame, Ind., 1980), pp. 145–158, Progr. Math., 14, Birkhauser, Boston, Mass., 1981. A. T. Huckleberry and M. Sauer, On the order of the automorphism group of a surface of general type. Math. Z. 205 (1990), no. 2, 321–329. Y. Kawamata, A generalization of Kodaira-Ramanujam’s vanishing theorem. Math. Ann. 261 (1982), no. 1, 43–46. Y. Kawamata, Kodaira dimension of algebraic fiber spaces over curves. Invent. Math. 66 (1982), no. 1, 57–71. Y. Kawamata, K. Matsuda and K. Matsuki, Introduction to the minimal model problem. Algebraic geometry, Sendai, 1985, 283–360, Adv. Stud. Pure Math., 10, North-Holland, Amsterdam, 1987. J. Kollar, Higher direct images of dualizing sheaves. I. Ann. of Math. (2) 123 (1986), no. 1, 11–42. J. Kollar, Singularities of pairs. Algebraic geometry—Santa Cruz 1995, 221–287, Proc. Sympos. Pure Math., 62, Part 1, Amer. Math. Soc., Providence, RI, 1997. J. Kollar and S. Mori, Birational geometry of algebraic varieties. With the collaboration of C. H. Clemens and A. Corti. Cambridge Tracts in Mathematics, 134. Cambridge University Press, Cambridge, 1998. S. J. Kovacs, Number of automorphisms of principally polarized abelian varieties. Advances in algebraic geometry motivated by physics (Lowell, MA, 2000), 3–7, Contemp. Math., 276, Amer. Math. Soc., Providence, RI, 2001. R. Lazarsfeld, Positivity in algebraic geometry. I; II. Classical setting: line bundles and linear series. Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics, 48; 49. Springer-Verlag, Berlin, 2004. S. Lee, Quartic-canonical systems on canonical threefolds of index 1. Special issue in honor of Robin Hartshorne. Comm. Algebra 28 (2000), no. 12, 5517–5530. Y. Miyaoka, The Chern classes and Kodaira dimension of a minimal variety. Algebraic geometry, Sendai, 1985, 449–476, Adv. Stud. Pure Math., 10, North-Holland, Amsterdam, 1987. S. Mori, Classification of higher-dimensional varieties. Algebraic geometry, Bowdoin, 1985, 269–331, Proc. Sympos. Pure Math., 46, Part 1, Amer. Math. Soc., Providence, RI, 1987. Z. Ran, The structure of Gauss-like maps. Compositio Math. 52 (1984), no. 2, 171–177. M. Reid, Young person’s guide to canonical singularities. Algebraic geometry, Bowdoin, 1985, 345–414, Proc. Sympos. Pure Math., 46, Part 1, Amer. Math. Soc., Providence, RI, 1987. E. Szabo, Bounding automorphism groups. Math. Ann. 304 (1996), no. 4, 801–811. K. Ueno, Classification theory of algebraic varieties and compact complex spaces. Notes written in collaboration with P. Cherenack. Lecture Notes in Mathematics, Vol. 439. Springer-Verlag, Berlin-New York, 1975. K. Ueno, A remark on automorphisms of Enriques surfaces. J. Fac. Sci. Univ. Tokyo Sect. I A Math. 23 (1976), no. 1, 149–165. E. Viehweg, Vanishing theorems. J. Reine Angew. Math. 335 (1982), 1–8. E. Viehweg, Die Additivitat der Kodaira Dimension fur projektive Faserraume uber Varietaten des allgemeinen Typs. J. Reine Angew. Math. 330 (1982), 132–142. B. Weisfeiler, Post-classification version of Jordan’s theorem on finite linear groups. Proc. Nat. Acad. Sci. U.S.A. 81 (1984), no. 16, Phys. Sci., 5278–5279. G. Xiao, On abelian automorphism group of a surface of general type. Invent. Math. 102 (1990), no. 3, 619–631. G. Xiao, Bound of automorphisms of surfaces of general type. I. Ann. of Math. (2) 139 (1994), no. 1, 51–77. G. Xiao, Bound of automorphisms of surfaces of general type. II. J. Algebraic Geom. 4 (1995), no. 4, 701–793. G. Xiao, Linear bound for abelian automorphisms of varieties of general type. J. Reine Angew. Math. 476 (1996), 201–207. S. T. Yau, On the Ricci curvature of a compact Kahler manifold and the complex Monge-Ampère equation. I. Comm. Pure Appl. Math. 31 (1978), no. 3, 339–411.
Department of Mathematics, National University of Singapore
2 Science Drive 2, Singapore 117543, SINGAPORE
E-mail: matzdq@nus.edu.sg
**Appendix**
============
$$\text{\bf A PRODUCT FORMULA FOR VOLUMES OF VARIETIES}$$
$$\text{By Yujiro Kawamata}$$
The volume $v(X)$ of a smooth projective variety $X$ is defined by $$v(X) = \text{lim sup} \frac{\dim H^0(X, mK_X)}{m^d/d!}$$ where $d = \dim X$. This is a birational invariant.
Let $f: X \to Y$ be a surjective morphism of smooth projective varieties with connected fibers. Assume that both $Y$ and the general fiber $F$ of $f$ are varieties of general type. Then $$\frac{v(X)}{d_X!} \ge \frac{v(Y)}{d_Y!}\frac{v(F)}{d_F!}$$ where $d_X = \dim X$, $d_Y = \dim Y$ and $d_F=\dim F$.
Let $H$ be an ample divisor on $Y$. There exists a positive integer $m_0$ such that $m_0K_Y - H$ is effective.
Let $\epsilon$ be a positive integer. By Fujita’s approximation theorem ([@F]), after replacing a birational model of $X$, there exists a positive integer $m_1$ and ample divisors $L$ on $F$ such that $m_1K_F - L$ is effective and $v(\frac 1{m_1}L) > v(F)- \epsilon$.
By Viehweg’s weak positivity theorem ([@V]), there exists a positive integer $k$ such that $S^k(f_*\mathcal{O}_X(m_1K_{X/Y}) \otimes \mathcal{O}_Y(H))$ is generically generated by global sections for a positive integer $k$. $k$ is a function on $H$ and $m_1$.
We have $$\begin{split}
&\text{rank Im}(S^mS^k(f_*\mathcal{O}_X(m_1K_{X/Y})) \to
f_*\mathcal{O}_X(km_1mK_{X/Y})) \\
&\ge \dim H^0(F,kmL) \\
&\ge (v(F)- 2 \epsilon)\frac{(km_1m)^{d_F}}{d_F!}
\end{split}$$ for sufficiently large $m$.
Then $$\begin{split}
&\dim H^0(X,km_1mK_X) \\
&\ge \dim H^0(Y,k(m_1-m_0)mK_Y) \times
(v(F)- 2 \epsilon)\frac{(km_1m)^{d_F}}{d_F!} \\
&\ge (v(Y) - \epsilon)\frac{(k(m_1-m_0)m)^{d_Y}}{d_Y!}
(v(F)- 2 \epsilon)\frac{(km_1m)^{d_F}}{d_F!} \\
&\ge (v(Y) - 2 \epsilon)(v(F)- 2 \epsilon)\frac{(km_1m)^{d_X}}{d_Y!d_F!}
\end{split}$$ if we take $m_1$ large compared with $m_0$ such that $$\frac{(v(Y) - \epsilon)}{(v(Y)- 2 \epsilon)} \ge (\frac{m_1}{m_1-m_0})^{d_Y}.$$
If $X = Y \times F$, then we have an equality in the formula. We expect that the equality implies the isotriviality of the family.
[F]{} Fujita, Takao. [*Approximating Zariski decomposition of big line bundles*]{}. Kodai Math. J. 17 (1994), no. 1, 1–3.
Viehweg, Eckart. [*Weak positivity and the additivity of the Kodaira dimension for certain fibre spaces*]{}. Algebraic varieties and analytic varieties (Tokyo, 1981), 329–353, Adv. Stud. Pure Math., 1, North-Holland, Amsterdam, 1983.
Department of Mathematical Sciences, University of Tokyo
Komaba, Meguro, Tokyo, 153-8914, JAPAN
E-mail: kawamata@ms.u-tokyo.ac.jp
|
/**
* Adds legend functionality to charts.
*
* @module charts
* @submodule charts-legend
*/
var DOCUMENT = Y.config.doc,
TOP = "top",
RIGHT = "right",
BOTTOM = "bottom",
LEFT = "left",
EXTERNAL = "external",
HORIZONTAL = "horizontal",
VERTICAL = "vertical",
WIDTH = "width",
HEIGHT = "height",
POSITION = "position",
_X = "x",
_Y = "y",
PX = "px",
LEGEND = {
setter: function(val)
{
var legend = this.get("legend");
if(legend)
{
legend.destroy(true);
}
if(val instanceof Y.ChartLegend)
{
legend = val;
legend.set("chart", this);
}
else
{
val.chart = this;
if(!val.hasOwnProperty("render"))
{
val.render = this.get("contentBox");
val.includeInChartLayout = true;
}
legend = new Y.ChartLegend(val);
}
return legend;
}
},
/**
* Contains methods for displaying items horizontally in a legend.
*
* @module charts
* @submodule charts-legend
* @class HorizontalLegendLayout
*/
HorizontalLegendLayout = {
/**
* Displays items horizontally in a legend.
*
* @method _positionLegendItems
* @param {Array} items Array of items to display in the legend.
* @param {Number} maxWidth The width of the largest item in the legend.
* @param {Number} maxHeight The height of the largest item in the legend.
* @param {Number} totalWidth The total width of all items in a legend.
* @param {Number} totalHeight The total height of all items in a legend.
* @param {Number} padding The left, top, right and bottom padding properties for the legend.
* @param {Number} horizontalGap The horizontal distance between items in a legend.
* @param {Number} verticalGap The vertical distance between items in a legend.
* @param {String} hAlign The horizontal alignment of the legend.
* @param {String} vAlign The vertical alignment of the legend.
* @protected
*/
_positionLegendItems: function(items, maxWidth, maxHeight, totalWidth, totalHeight, padding, horizontalGap, verticalGap, hAlign, vAlign)
{
var i = 0,
rowIterator = 0,
item,
node,
itemWidth,
itemHeight,
len,
width = this.get("width"),
rows,
rowsLen,
row,
totalWidthArray,
legendWidth,
topHeight = padding.top - verticalGap,
limit = width - (padding.left + padding.right),
left,
top,
right,
bottom;
HorizontalLegendLayout._setRowArrays(items, limit, horizontalGap);
rows = HorizontalLegendLayout.rowArray;
totalWidthArray = HorizontalLegendLayout.totalWidthArray;
rowsLen = rows.length;
for(; rowIterator < rowsLen; ++ rowIterator)
{
topHeight += verticalGap;
row = rows[rowIterator];
len = row.length;
legendWidth = HorizontalLegendLayout.getStartPoint(width, totalWidthArray[rowIterator], hAlign, padding);
for(i = 0; i < len; ++i)
{
item = row[i];
node = item.node;
itemWidth = item.width;
itemHeight = item.height;
item.x = legendWidth;
item.y = 0;
left = !isNaN(left) ? Math.min(left, legendWidth) : legendWidth;
top = !isNaN(top) ? Math.min(top, topHeight) : topHeight;
right = !isNaN(right) ? Math.max(legendWidth + itemWidth, right) : legendWidth + itemWidth;
bottom = !isNaN(bottom) ? Math.max(topHeight + itemHeight, bottom) : topHeight + itemHeight;
node.setStyle("left", legendWidth + PX);
node.setStyle("top", topHeight + PX);
legendWidth += itemWidth + horizontalGap;
}
topHeight += item.height;
}
this._contentRect = {
left: left,
top: top,
right: right,
bottom: bottom
};
if(this.get("includeInChartLayout"))
{
this.set("height", topHeight + padding.bottom);
}
},
/**
* Creates row and total width arrays used for displaying multiple rows of
* legend items based on the items, available width and horizontalGap for the legend.
*
* @method _setRowArrays
* @param {Array} items Array of legend items to display in a legend.
* @param {Number} limit Total available width for displaying items in a legend.
* @param {Number} horizontalGap Horizontal distance between items in a legend.
* @protected
*/
_setRowArrays: function(items, limit, horizontalGap)
{
var item = items[0],
rowArray = [[item]],
i = 1,
rowIterator = 0,
len = items.length,
totalWidth = item.width,
itemWidth,
totalWidthArray = [[totalWidth]];
for(; i < len; ++i)
{
item = items[i];
itemWidth = item.width;
if((totalWidth + horizontalGap + itemWidth) <= limit)
{
totalWidth += horizontalGap + itemWidth;
rowArray[rowIterator].push(item);
}
else
{
totalWidth = horizontalGap + itemWidth;
if(rowArray[rowIterator])
{
rowIterator += 1;
}
rowArray[rowIterator] = [item];
}
totalWidthArray[rowIterator] = totalWidth;
}
HorizontalLegendLayout.rowArray = rowArray;
HorizontalLegendLayout.totalWidthArray = totalWidthArray;
},
/**
* Returns the starting x-coordinate for a row of legend items.
*
* @method getStartPoint
* @param {Number} w Width of the legend.
* @param {Number} totalWidth Total width of all labels in the row.
* @param {String} align Horizontal alignment of items for the legend.
* @param {Object} padding Object contain left, top, right and bottom padding properties.
* @return Number
* @protected
*/
getStartPoint: function(w, totalWidth, align, padding)
{
var startPoint;
switch(align)
{
case LEFT :
startPoint = padding.left;
break;
case "center" :
startPoint = (w - totalWidth) * 0.5;
break;
case RIGHT :
startPoint = w - totalWidth - padding.right;
break;
}
return startPoint;
}
},
/**
* Contains methods for displaying items vertically in a legend.
*
* @module charts
* @submodule charts-legend
* @class VerticalLegendLayout
*/
VerticalLegendLayout = {
/**
* Displays items vertically in a legend.
*
* @method _positionLegendItems
* @param {Array} items Array of items to display in the legend.
* @param {Number} maxWidth The width of the largest item in the legend.
* @param {Number} maxHeight The height of the largest item in the legend.
* @param {Number} totalWidth The total width of all items in a legend.
* @param {Number} totalHeight The total height of all items in a legend.
* @param {Number} padding The left, top, right and bottom padding properties for the legend.
* @param {Number} horizontalGap The horizontal distance between items in a legend.
* @param {Number} verticalGap The vertical distance between items in a legend.
* @param {String} hAlign The horizontal alignment of the legend.
* @param {String} vAlign The vertical alignment of the legend.
* @protected
*/
_positionLegendItems: function(items, maxWidth, maxHeight, totalWidth, totalHeight, padding, horizontalGap, verticalGap, hAlign, vAlign)
{
var i = 0,
columnIterator = 0,
item,
node,
itemHeight,
itemWidth,
len,
height = this.get("height"),
columns,
columnsLen,
column,
totalHeightArray,
legendHeight,
leftWidth = padding.left - horizontalGap,
legendWidth,
limit = height - (padding.top + padding.bottom),
left,
top,
right,
bottom;
VerticalLegendLayout._setColumnArrays(items, limit, verticalGap);
columns = VerticalLegendLayout.columnArray;
totalHeightArray = VerticalLegendLayout.totalHeightArray;
columnsLen = columns.length;
for(; columnIterator < columnsLen; ++ columnIterator)
{
leftWidth += horizontalGap;
column = columns[columnIterator];
len = column.length;
legendHeight = VerticalLegendLayout.getStartPoint(height, totalHeightArray[columnIterator], vAlign, padding);
legendWidth = 0;
for(i = 0; i < len; ++i)
{
item = column[i];
node = item.node;
itemHeight = item.height;
itemWidth = item.width;
item.y = legendHeight;
item.x = leftWidth;
left = !isNaN(left) ? Math.min(left, leftWidth) : leftWidth;
top = !isNaN(top) ? Math.min(top, legendHeight) : legendHeight;
right = !isNaN(right) ? Math.max(leftWidth + itemWidth, right) : leftWidth + itemWidth;
bottom = !isNaN(bottom) ? Math.max(legendHeight + itemHeight, bottom) : legendHeight + itemHeight;
node.setStyle("left", leftWidth + PX);
node.setStyle("top", legendHeight + PX);
legendHeight += itemHeight + verticalGap;
legendWidth = Math.max(legendWidth, item.width);
}
leftWidth += legendWidth;
}
this._contentRect = {
left: left,
top: top,
right: right,
bottom: bottom
};
if(this.get("includeInChartLayout"))
{
this.set("width", leftWidth + padding.right);
}
},
/**
* Creates column and total height arrays used for displaying multiple columns of
* legend items based on the items, available height and verticalGap for the legend.
*
* @method _setColumnArrays
* @param {Array} items Array of legend items to display in a legend.
* @param {Number} limit Total available height for displaying items in a legend.
* @param {Number} verticalGap Vertical distance between items in a legend.
* @protected
*/
_setColumnArrays: function(items, limit, verticalGap)
{
var item = items[0],
columnArray = [[item]],
i = 1,
columnIterator = 0,
len = items.length,
totalHeight = item.height,
itemHeight,
totalHeightArray = [[totalHeight]];
for(; i < len; ++i)
{
item = items[i];
itemHeight = item.height;
if((totalHeight + verticalGap + itemHeight) <= limit)
{
totalHeight += verticalGap + itemHeight;
columnArray[columnIterator].push(item);
}
else
{
totalHeight = verticalGap + itemHeight;
if(columnArray[columnIterator])
{
columnIterator += 1;
}
columnArray[columnIterator] = [item];
}
totalHeightArray[columnIterator] = totalHeight;
}
VerticalLegendLayout.columnArray = columnArray;
VerticalLegendLayout.totalHeightArray = totalHeightArray;
},
/**
* Returns the starting y-coordinate for a column of legend items.
*
* @method getStartPoint
* @param {Number} h Height of the legend.
* @param {Number} totalHeight Total height of all labels in the column.
* @param {String} align Vertical alignment of items for the legend.
* @param {Object} padding Object contain left, top, right and bottom padding properties.
* @return Number
* @protected
*/
getStartPoint: function(h, totalHeight, align, padding)
{
var startPoint;
switch(align)
{
case TOP :
startPoint = padding.top;
break;
case "middle" :
startPoint = (h - totalHeight) * 0.5;
break;
case BOTTOM :
startPoint = h - totalHeight - padding.bottom;
break;
}
return startPoint;
}
},
CartesianChartLegend = Y.Base.create("cartesianChartLegend", Y.CartesianChart, [], {
/**
* Redraws and position all the components of the chart instance.
*
* @method _redraw
* @private
*/
_redraw: function()
{
if(this._drawing)
{
this._callLater = true;
return;
}
this._drawing = true;
this._callLater = false;
var w = this.get("width"),
h = this.get("height"),
layoutBoxDimensions = this._getLayoutBoxDimensions(),
leftPaneWidth = layoutBoxDimensions.left,
rightPaneWidth = layoutBoxDimensions.right,
topPaneHeight = layoutBoxDimensions.top,
bottomPaneHeight = layoutBoxDimensions.bottom,
leftAxesCollection = this.get("leftAxesCollection"),
rightAxesCollection = this.get("rightAxesCollection"),
topAxesCollection = this.get("topAxesCollection"),
bottomAxesCollection = this.get("bottomAxesCollection"),
i = 0,
l,
axis,
graphOverflow = "visible",
graph = this.get("graph"),
topOverflow,
bottomOverflow,
leftOverflow,
rightOverflow,
graphWidth,
graphHeight,
graphX,
graphY,
allowContentOverflow = this.get("allowContentOverflow"),
diff,
rightAxesXCoords,
leftAxesXCoords,
topAxesYCoords,
bottomAxesYCoords,
legend = this.get("legend"),
graphRect = {};
if(leftAxesCollection)
{
leftAxesXCoords = [];
l = leftAxesCollection.length;
for(i = l - 1; i > -1; --i)
{
leftAxesXCoords.unshift(leftPaneWidth);
leftPaneWidth += leftAxesCollection[i].get("width");
}
}
if(rightAxesCollection)
{
rightAxesXCoords = [];
l = rightAxesCollection.length;
i = 0;
for(i = l - 1; i > -1; --i)
{
rightPaneWidth += rightAxesCollection[i].get("width");
rightAxesXCoords.unshift(w - rightPaneWidth);
}
}
if(topAxesCollection)
{
topAxesYCoords = [];
l = topAxesCollection.length;
for(i = l - 1; i > -1; --i)
{
topAxesYCoords.unshift(topPaneHeight);
topPaneHeight += topAxesCollection[i].get("height");
}
}
if(bottomAxesCollection)
{
bottomAxesYCoords = [];
l = bottomAxesCollection.length;
for(i = l - 1; i > -1; --i)
{
bottomPaneHeight += bottomAxesCollection[i].get("height");
bottomAxesYCoords.unshift(h - bottomPaneHeight);
}
}
graphWidth = w - (leftPaneWidth + rightPaneWidth);
graphHeight = h - (bottomPaneHeight + topPaneHeight);
graphRect.left = leftPaneWidth;
graphRect.top = topPaneHeight;
graphRect.bottom = h - bottomPaneHeight;
graphRect.right = w - rightPaneWidth;
if(!allowContentOverflow)
{
topOverflow = this._getTopOverflow(leftAxesCollection, rightAxesCollection);
bottomOverflow = this._getBottomOverflow(leftAxesCollection, rightAxesCollection);
leftOverflow = this._getLeftOverflow(bottomAxesCollection, topAxesCollection);
rightOverflow = this._getRightOverflow(bottomAxesCollection, topAxesCollection);
diff = topOverflow - topPaneHeight;
if(diff > 0)
{
graphRect.top = topOverflow;
if(topAxesYCoords)
{
i = 0;
l = topAxesYCoords.length;
for(; i < l; ++i)
{
topAxesYCoords[i] += diff;
}
}
}
diff = bottomOverflow - bottomPaneHeight;
if(diff > 0)
{
graphRect.bottom = h - bottomOverflow;
if(bottomAxesYCoords)
{
i = 0;
l = bottomAxesYCoords.length;
for(; i < l; ++i)
{
bottomAxesYCoords[i] -= diff;
}
}
}
diff = leftOverflow - leftPaneWidth;
if(diff > 0)
{
graphRect.left = leftOverflow;
if(leftAxesXCoords)
{
i = 0;
l = leftAxesXCoords.length;
for(; i < l; ++i)
{
leftAxesXCoords[i] += diff;
}
}
}
diff = rightOverflow - rightPaneWidth;
if(diff > 0)
{
graphRect.right = w - rightOverflow;
if(rightAxesXCoords)
{
i = 0;
l = rightAxesXCoords.length;
for(; i < l; ++i)
{
rightAxesXCoords[i] -= diff;
}
}
}
}
graphWidth = graphRect.right - graphRect.left;
graphHeight = graphRect.bottom - graphRect.top;
graphX = graphRect.left;
graphY = graphRect.top;
if(legend)
{
if(legend.get("includeInChartLayout"))
{
switch(legend.get("position"))
{
case "left" :
legend.set("y", graphY);
legend.set("height", graphHeight);
break;
case "top" :
legend.set("x", graphX);
legend.set("width", graphWidth);
break;
case "bottom" :
legend.set("x", graphX);
legend.set("width", graphWidth);
break;
case "right" :
legend.set("y", graphY);
legend.set("height", graphHeight);
break;
}
}
}
if(topAxesCollection)
{
l = topAxesCollection.length;
i = 0;
for(; i < l; i++)
{
axis = topAxesCollection[i];
if(axis.get("width") !== graphWidth)
{
axis.set("width", graphWidth);
}
axis.get("boundingBox").setStyle("left", graphX + PX);
axis.get("boundingBox").setStyle("top", topAxesYCoords[i] + PX);
}
if(axis._hasDataOverflow())
{
graphOverflow = "hidden";
}
}
if(bottomAxesCollection)
{
l = bottomAxesCollection.length;
i = 0;
for(; i < l; i++)
{
axis = bottomAxesCollection[i];
if(axis.get("width") !== graphWidth)
{
axis.set("width", graphWidth);
}
axis.get("boundingBox").setStyle("left", graphX + PX);
axis.get("boundingBox").setStyle("top", bottomAxesYCoords[i] + PX);
}
if(axis._hasDataOverflow())
{
graphOverflow = "hidden";
}
}
if(leftAxesCollection)
{
l = leftAxesCollection.length;
i = 0;
for(; i < l; ++i)
{
axis = leftAxesCollection[i];
axis.get("boundingBox").setStyle("top", graphY + PX);
axis.get("boundingBox").setStyle("left", leftAxesXCoords[i] + PX);
if(axis.get("height") !== graphHeight)
{
axis.set("height", graphHeight);
}
}
if(axis._hasDataOverflow())
{
graphOverflow = "hidden";
}
}
if(rightAxesCollection)
{
l = rightAxesCollection.length;
i = 0;
for(; i < l; ++i)
{
axis = rightAxesCollection[i];
axis.get("boundingBox").setStyle("top", graphY + PX);
axis.get("boundingBox").setStyle("left", rightAxesXCoords[i] + PX);
if(axis.get("height") !== graphHeight)
{
axis.set("height", graphHeight);
}
}
if(axis._hasDataOverflow())
{
graphOverflow = "hidden";
}
}
this._drawing = false;
if(this._callLater)
{
this._redraw();
return;
}
if(graph)
{
graph.get("boundingBox").setStyle("left", graphX + PX);
graph.get("boundingBox").setStyle("top", graphY + PX);
graph.set("width", graphWidth);
graph.set("height", graphHeight);
graph.get("boundingBox").setStyle("overflow", graphOverflow);
}
if(this._overlay)
{
this._overlay.setStyle("left", graphX + PX);
this._overlay.setStyle("top", graphY + PX);
this._overlay.setStyle("width", graphWidth + PX);
this._overlay.setStyle("height", graphHeight + PX);
}
},
/**
* Positions the legend in a chart and returns the properties of the legend to be used in the
* chart's layout algorithm.
*
* @method _getLayoutDimensions
* @return {Object} The left, top, right and bottom values for the legend.
* @protected
*/
_getLayoutBoxDimensions: function()
{
var box = {
top: 0,
right: 0,
bottom: 0,
left: 0
},
legend = this.get("legend"),
position,
direction,
dimension,
size,
w = this.get(WIDTH),
h = this.get(HEIGHT),
gap;
if(legend && legend.get("includeInChartLayout"))
{
gap = legend.get("styles").gap;
position = legend.get(POSITION);
if(position != EXTERNAL)
{
direction = legend.get("direction");
dimension = direction == HORIZONTAL ? HEIGHT : WIDTH;
size = legend.get(dimension);
box[position] = size + gap;
switch(position)
{
case TOP :
legend.set(_Y, 0);
break;
case BOTTOM :
legend.set(_Y, h - size);
break;
case RIGHT :
legend.set(_X, w - size);
break;
case LEFT:
legend.set(_X, 0);
break;
}
}
}
return box;
},
/**
* Destructor implementation for the CartesianChart class. Calls destroy on all axes, series, legend (if available) and the Graph instance.
* Removes the tooltip and overlay HTML elements.
*
* @method destructor
* @protected
*/
destructor: function()
{
var legend = this.get("legend");
if(legend)
{
legend.destroy(true);
}
}
}, {
ATTRS: {
legend: LEGEND
}
});
Y.CartesianChart = CartesianChartLegend;
|
Ms. Maxwell had negligible experience as an environmental activist. Her preferred method of oceangoing was aboard a luxury yacht, which, according to Mr. Mason, was for her the pre-eminent symbol of “status and freedom.” It was through boating that she drew her inspiration for the foundation.
She spent much of her time in the late 1980s on the Lady Ghislaine, a nearly 200-foot boat owned by her father, the media mogul Robert Maxwell. It had a Jacuzzi, a sauna, a gym and private disco. Deep in debt, he bilked the pensions of thousands of his employees, and his body was discovered in the ocean off the Canary Islands, where he had taken the Lady Ghislaine in 1991.
The death was ruled an accident. The family reportedly lost almost everything, including the boat.
Ms. Maxwell, then living in New York, became known for her romantic relationship with Mr. Epstein, who was an all-purpose adviser for the billionaire Leslie Wexner. (Mr. Wexner said in a letter to the Wexner Foundation that, in 2007, he discovered misappropriation of his funds by Mr. Epstein.)
One of Mr. Epstein’s duties was handling contracts for the Limitless, a mammoth yacht bought by Mr. Wexner and designed by Bannenberg & Rowell. Ms. Maxwell was eager to get aboard when it was finished but never did, according to Craig Tafoya, its former captain.
“Ghislaine would always call me and say, ‘I’m coming down to use the boat with some friends. I would always tell her, ‘I have to call the owner. I can’t just let you on the boat.’ And she would never show up,” said Mr. Tafoya, who took this to mean that she never got permission. “She did that half a dozen times. And in talking to a guy who worked for Bannenberg, he said, ‘she does that all the time. She does it when she’s in front of all her girlfriends and wants to brag that she can go use someone’s yacht.’”
Her Plan B |
By continuing to use this site you consent to the use of cookies on your device as described in our
Cookie Policy unless you have disabled them. You can change your Cookie Settings at any time but parts of our site will not function correctly without them.
Speaking to ANI, Swamy said that the government's stand on homosexuality was tougher than his and that while he respected the right to privacy, the government left it to the top court to decide.
"The government's affidavit stated that the court was to decide whether homosexual activity in private is a criminal act or not. The government's stand is harder than my stand, I said clearly on privacy, no question. Right to Privacy is protected. But they (the government) are saying the court should now decide," Swamy told ANI.
He also said that it was the government's view that homosexuality was an "unnatural act" and a genetic flaw.
"It is a handicap, for which people should not be discriminated against, but at the same time you can't celebrate it," Swamy said.
Earlier, on Wednesday, Swamy had stated that homosexuality was a danger to national security and that if the Supreme Court ruled it to be normal and a free choice, the government should constitute a seven-judge or a nine-judge bench to review it.
In May, the apex court decided to hear the plea filed by the Indian Institute of Technology's LGBT alumni association, seeking to scrap Section 377, that criminalises homosexuality.
In 2009, the Delhi High Court had decriminalised Section 377, but the order was later overturned by a Supreme-Court bench in 2013.
Categorised as an unnatural offence, consensual intercourse between persons of same-sex is termed 'against the order of nature' under Section 377, and can be punishable by life imprisonment.
Last year, the Supreme Court ruled that individual privacy is a fundamental right that cannot be denied "even if a minuscule fraction of population is affected". "Privacy includes at its core the preservation of personal intimacies, the sanctity of family life, marriage, procreation, the home and sexual orientation... Privacy also connotes a right to be left alone." |
Welcome to the newest edition of BGN Radio! BGN Radio is Bleeding Green Nation's regularly scheduled podcast. Each week some combination of Brandon Lee Gowton, Dave Mangels, Adam Hermann, James Seltzer, Patrick Wall, Matt Dering, Ben Natan, and host John Barchard will be here to talk about the Philadelphia Eagles. This episode is brought to you by:
Powered by @ClippitTV and @SportsRadioWIP
Brought to you by Clippit - Watch TV. Make Clips. Share. (www.clippit.tv).
Make 30 second, full HD clips from live television. App Store | Android
This episode is also brought to you by the Tapped Beer Festival happening in Norristown, PA! Over 100 beers, wines and ciders to taste from Victory, Sly Fox, Troegs and much much more! Grab your tickets NOW for early entry at tappedfest.com/philly
SHOW TOPICS
Yes Chicago is bad team, no it doesn’t take away from Wentz’s performance.
It was 9-7 on the scoreboard for a good chunk of time but the Eagles dominated the Bears early in this one.
Big Balls Doug? Another impressive and aggressive gameplan from Mr. Pederson.
Ryan Mathews showed us again why he went in the first round. Please stay healthy.
The Eagles for the 2nd week in a row have knocked out a starting a QB out of the game.
Agholor & Matthews old ways popped up again.
Carson Wentz was once again in command, directing pre-snap traffic and looked like a veteran once again.
Jalen Mills looked liked a rookie early, but played a solid game.
When this team gets a lead, the defense turns into a bunch of animals.
Eagles are 2-0 with each win being 14+ points. The last 13 teams to start 2-0 with a pair of wins by 14+ went to the playoffs.
Final Thoughts & more!
Join us again this Saturday from 4p-6p EST on 94WIP!
HOW TO LISTEN
STREAM:
DOWNLOAD: Click here to download this episode.
In order to grow BGN Radio even more, we'd like to ask a few favors of you:
• Click here to SUBSCRIBE to the official BGN Radio iTunes feed. Also please rate the show (5 stars, obviously!) and leave a review. Not only do we love to hear your feedback, but this also helps other Eagles fans find the show.
• Are you interested in advertising with BGN Radio? Contact us at bgnradiopod@gmail.com.
• Check out our website - www.BGNRadio.com. Go buy apparel from the BGN Radio Shop!
• Speaking of helping others find the show, don't be shy to share the podcast with your friends. They'll love it, we promise.
• Follow BGN Radio on Twitter: @BGN_Radio. (If you like fantasy football, also check out @BGN_TWIF.)
• Follow BGN Radio on SoundCloud.
• Rate BGN Radio on Stitcher and leave a review.
• Listen to BGN Radio on Google Play. |
News from the Hill
On Dec. 22, 2017, President Trump signed the tax reform bill into law—with all changes under this legislation set to take effect on Jan. 1 (see here for a summary of how this legislation impacts philanthropy). Though the overhaul has become the law of the land, Congress will likely still pass an additional bill to correct technical errors that appear in the legislation—which will require 60 votes to pass in the Senate, and therefore, require Democratic support. Chairman of the House Ways and Means Committee Kevin Brady (R-TX) has indicated his preference to limit the bill to make only necessary changes, but Democratic leaders have suggested that they may push back.
The other big tax news of the week was that Senate Finance Committee Chairman Orrin Hatch (R-UT) will retire from Congress at the end of 2018—concluding his career as the longest-serving Republican senator. His departure opens the powerful chairmanship to another Republican member of the Finance Committee. Likely candidates include Sen. Chuck Grassley (R-IA), who will become the most senior Republican member of the committee once Sen. Hatch leaves, with Sen. Mike Crapo (R-ID) in line behind him.
Deadline to Pass FY ‘18 Spending Measure Looms
On Dec. 22, Congress passed a temporary stop-gap measure that funds the government through Jan. 19 and avoids a shutdown. As the Senate returned from their holiday recess yesterday and the House returns next week, House and Senate leaders face a tight deadline to prevent the government from grinding to a halt. Yesterday afternoon, the four congressional leaders—House Speaker Paul Ryan (R-WI), House Minority Leader Nancy Pelosi (D-CA), Senate Majority Leader Mitch McConnell (R-KY), and Senate Minority Leader Chuck Schumer (D-NY)—met with representatives from the White House to begin negotiations on a funding bill for the rest of fiscal year (FY) 2018.
According to The Wall Street Journal, “Aides to the top congressional leaders have been meeting with White House staff including Mr. [Marc] Short [White House director of legislative affairs] for weeks, but GOP leaders were focused on passing the tax overhaul, which squeaked through both chambers late last month. Now with the tax bill out of the way, aides said they were optimistic lawmakers could get closer to striking a budget deal, though sticking points remain.”
The potential sticking points include: raising the sequester budget spending caps that are set to kick-in due to a 2011 budget deal—(Congress has reached multiple deals since 2011 to avert the across-the-board spending caps from being triggered); legal protections for children who were brought undocumented into the U.S. that had been protected under the Deferred Action for Childhood Arrivals program (DACA), which was ended in September by President Donald Trump; and a reauthorization of the recently expired Children’s Health Insurance Program (CHIP).
Two New Senators Sworn In
Yesterday, two new Democratic senators were sworn in—narrowing the Republican majority to just one seat (51-49). Sen. Doug Jones (D-AL) replaces former Sen. “Big” Luther Strange (R-AL) after defeating Roy Moore in a special election last month. Sen. Strange, who was appointed as an interim senator when Jeff Sessions became the U.S. Attorney General, lost to Moore in a tight primary. According to NPR, “The upset victory by Jones narrowed Republicans' hold on the Senate to only one seat, 51-49. That's noteworthy going into a midterm election year, making Democrats' impossible path to a Senate majority now just improbable. They're still largely on defense, with 10 Senate Democrats up for re-election in states that President Trump won. But if they can flip the GOP-held Arizona and Nevada seats—which are highly competitive—and defend all their incumbents, there's a way.”
The day also saw the former Minnesota Lieutenant-Governor Tina Smith sworn in. Sen. Smith was appointed to replace Sen. Al Franken (D-MN), who announced his resignation last month after multiple allegations of sexual harassment. NPR also noted that, “Even though Jones won a special election for his seat and Smith was appointed, it's Smith who will have the seniority edge since she and Jones are being sworn in at the same time—simply because Minnesota is more populous than Alabama.”
With the addition of another Democrat in the Senate (Jones replacing Strange), there may be some tweaks to committee assignments. According to POLITICO, “Some logistical issues will also have to be worked out with a narrower Republican margin in the Senate. For instance, McConnell and Schumer are still discussing whether the number of seats allotted to each party on committees will be adjusted to reflect the smaller GOP majority, aides said.”
Happening in the States
State Legislatures Convene to Confront Uncertainty Caused by the Federal Tax Bill
January marks the beginning of legislative sessions in thirty-eight states, and the impact of the federal tax bill enacted just two weeks ago is top of mind for governors and legislators. Forty states link their tax laws to the federal code, yet the changes will affect states differently and require varied responses for legislatures. For instance, the standard deduction in 12 states automatically went up on January 1 because their tax codes follow the federal standard deduction; eight states match the federal personal exemption, which has now been repealed.
While a few states, such as Colorado and Maryland, may initially see increased revenues due to how their tax laws interact with the Internal Revenue Code, many more are predicting losses in tax receipts that could affect their budgets for the current fiscal year that typically ends on June 30. Any one change to federal tax law can knock state budgets out of alignment and the states are in no condition to absorb revenue losses. In 2017, 22 states suffered revenue shortfalls, making them unable to maintain services at existing levels. As regular readers of Washington Snapshot have seen, the ongoing Illinois budget crisis is legendary, while New Mexico and North Dakota are mired in at least their third straight year of spending cuts. Kentucky has a $1 billion budget hole and is already cutting human services, and Iowa expects to make cuts of between $45 - $90 million by June 30. Oklahoma doesn’t even have a budget for the Fiscal Year that started July 2017.
Growing budget deficits and the need to open up state tax laws make clear that the 2018 legislative session will be challenging for nonprofits and foundations. In “Nonprofits Must Move Swiftly to Fight for Sound Public Policies,” published yesterday in the Chronicle of Philanthropy, Tim Delaney and David L. Thompson of the National Council of Nonprofits translate these challenges into likely policy actions in the states. The article predicts attempts will be made to reconfigure state and local tax laws in ways that lead to new levies on tax-exempt entities, such as additional fees and payments in lieu of taxes (PILOTs) on nonprofit-owned real estate, penalties on nonprofit salaries, and excise taxes on some endowments. Further, as governments at all levels are forced to cut spending, more work will likely fall on nonprofits to help people hurt by the spending reductions. The article explains, “Expect nonprofits to have to seek more money from foundations to cover those costs — think of it as a new tax on philanthropy to subsidize decisions of politicians.” |
Good to be Antibad
Photography by David Reiss
Fashion Director and Interview by Ursula Lake
Make magazine chats to Agatha Lintott the woman responsible for Antibad an innovative new website that curates and champions brands that are as chic as they are environmentally sound. The first ‘green’ online boutique.
Q. How did you get started with Antibad?
A. I had been working for Tom Ford and then Burberry, and when I finished at Burberry my original idea had been to start my own brand. As I am half Peruvian and I have family there, initially, I wanted to create something with artisans in Peru, but I was mindful that I didn’t want it to be too ethnic, I wanted to create something that my old colleagues at Burberry would want to wear to work. I went out to Peru but realised quickly that it wasn’t really going to work! I concluded that one of my strengths was curating brands and over time managed to find a bunch of them with similar mindsets about sustainability and environmental concerns and decided to bring them together, so thus Antibad was created. And the name comes from the fact that we are anti bad fashion!
The idea eventually is to be selling the brand as a marketplace structure, which is already beginning to come together as brands are approaching me to sell their labels.
Q. To keep with the concept of sustainability are you working with seasons or using a more trans-seasonal approach to what you are selling?
A. Absolutely, I only have a few brands that work within the traditional parameters of ‘seasons’ or go on sale. We don’t go on sale or use the commonly used trick of ‘black Friday sales’ as I think it’s really damaging to the brands. We believe in a fair price for the intrinsic craftsmanship of what we are selling. The idea of sales makes people consume fashion in a less responsible way as they often buy things because they are cheap and not because they really love them or they have a valid place in their wardrobe. Also, fashion does work cyclically. There might be a time when you don’t want to wear something but I think if you have bought consciously you will probably feel inspired to wear it in the future again. Women on average only wear 20 percent of their wardrobe and I think that is a great shame and something we should be addressing.
Q. How do you quantify the ‘green’ attributes that a designer needs to be adhering to in order for it to make it on the Antibad website?
A. Fair-trade should be the requirement but it’s an expensive process to be listed legally under this certification, therefore most of the designers I work with do so in the way that would make them officially fair-trade. This means that they are small and they know every single person who makes their clothes by name. They pay fair wages, they have a culture of respect for their workers. The other side of it is the materials used: the dyes, the fabrics, whether they are using dead-stock fabrics. There are no set rules about what makes a brand ethical exactly and there are compromises. What I am looking for, above all, is transparency about all of their working practices so that one can give the customer the choice about how they want to invest their money in fashion and what aspects are important to them. Do they want to buy locally? Do they want to buy recycled fabrics? Does it need to be vegan? That is why we shop by a filter on the site so that the customer can narrow down their choices by the parameters which are closest to their hearts in terms of environmental concerns.
The minimum for the site is that the workers are treated within the rules of fair-trade, there are no virgin polyesters or toxic dyes used and it has to be as close to carbon neutral as is possible. It is difficult and sometimes it feels overwhelming as customers feel like they can’t do anything ‘right’ in terms of ticking all the boxes for ethical purchases. An important reason why people respond to the site as the consumer is able to show with their money what ethical cause they want to support.
Q. For you, what are the biggest issues surrounding ethical fashion?
A. People and their treatment is one of my primary concerns. There are laws out there to protect workers in terms of minimum wage and slavery, but I don’t think that the government puts enough pressure on business to manage that properly and really invest in it. There are some very big companies with corporate social responsibility teams and they have sustainability teams. I met with someone who worked in one of those teams and she said that although they try their hardest, they don’t know one hundred percent if their products are ethical as the businesses are too large to police properly. I think that the government really need to do something to support big business to make sure that the products are ethical.
Q. What do you think about the way the media in general talks about fashion?
A. I think we need to re-address the way that it is spoken about in all areas of media. Change the subject from newness, and that is tricky as a lot of these business rely on consumption. However, I do think that something needs to be addressed when some fashion businesses are bringing out sixty collections a season! I am trying to address that as well as ultimately, I am a shop, and I do want people to purchase things! I am doing that by hoping that I am encouraging people to really think about what they are buying. I would like to encourage people to shop for quality rather than quantity. Even in the way we portray our products on the site for example, we don’t use models. We use people who are involved in activism in some way, so they truly represent the brand.
Q. You have vintage pieces on the site, how do you about finding those pieces? Do you source them personally?
A. There is a vintage wholesaler that I work with and I tend to buy everything from there. They find their vintage in Italy and then I curate special pieces for the site. It was always important for me to include vintage as I think it’s probably the most sustainable way to shop. They are often really special pieceshand made, that often go up in value as they are unique.
Q. Who or whatare the worst offenders in fashion?
A. Leather and denim are some ofthe worst polluters in the world: They use toxic dyes that flood into the water systems as well as consuming a lot of water; by and large, they are very unethical in their practices. The leather we have on the site is all vegetable tanned and a by-product of the food industry. We use denim from brands like Happy Haus, that are made from 100% washed denim that’s been DETOX certified and Greenpeace approved.
Q. Has all this awareness led you to more ‘green’ in other areas and practices of your life?
A. I am certainly trying. I have learned a lot from all the people I have met on this journey, soI try to use this knowledge in all areas of my life. Things like plastic are now very important to me, whereas before I don’t think I really thought too much about it. I am definitely a lot more conscious about what I am eating and where it comes from than I have ever been. Also, beauty products: I never used to think about them before at all, which looking back was really naive. I certainly think now about my own personal impact on the world, what goes down the sink and what goes in the bin?
I would love to have a shop where there was a washing station and also a mending station too. The sad things is that a great deal of ‘affordable’ fashion is that the quality is so bad that it is meant to go in the bin after a few washes. You are not going to mend it if you can buy another one for £2.00 and that is a really bad message. If we invested that little bit more, things would last that much longer clothes that will stay in your wardrobe for years and work really hard for you. Often, because we spent a bit more money on shoes or bags, we have them resoled or mended because we have spent that extra money but the same ethos often doesn’t spread to mending clothes and I would like to change that too.
Q. Do you hope at some point to do your own label?
A. I would love to do collaborations with brands and use their expertise and infrastructure to work with them to produce something ethical. I have already been talking to one of our designers on the site, Clan of Crows, about collaborating on a range of dresses that would be exclusively available on Antibad.
Q. Are there any new brands that coming to the site soon that we need to look out for?
I have a list of over 200 brands that I would love to work with, which is really exciting. There are some incredible brands out there, whichis reassuring, but I don’t want it to be an overload of products that is overwhelming to the customer. I want to keep it very curated. Fundamentally, I think that we should be able to treat ourselves. If you want a new dress, get a new dress! The main message that I strive to get across is one of buying better. Choosing the better alternatives from what is out there.
Q This is our British issue, so I was wondering how well you think Britain does in our environmental credentials?
In fashion, I would say we are not doing that well. It is nowhere near as big as places like California for instance or Germany and the Netherlands. We are not yet at a point where sustainable fashion is mixed in with normal fashion. In the media, we are still always lumped into featureson sustainability and these brands don’t appear in the fashion pages. Whereas in those other countries it’s being worn by influencers or celebrities and it is more commonplace. They are not necessarily buying something because it’s sustainable, it’s more that it is just really beautiful. I think we have a few years to go before we catch up.
There are some great brands starting up, but a big problem is that we don’t have the resources to make things ethically in the same way that they do. In some countries, they have the ability to grow, weave, dye and manufacture all in one country, which in itself is a way of being more ethical in clothing production. We are lacking that infrastructure in this country.
Q. Do you have any simple advice for anyone trying to be a bit more ethical in their fashion choices that might make it a bit easier for them?
A. Don’t become overwhelmed with trying to do everything at once. I have friends that come to me feeling guilty about their purchases all the time. I try to placate them and tell them that it is fine, but maybe next time, try to do their research. Spend a bit more time and don’t buy on impulse or on a whim. Don’t also feel that you have to throw away all your ‘bad fashion’ as that only compounds the problem. I have things in my wardrobe from Zara or H&M from years ago and I am not going to throw them away. But now I only buy sustainably and gradually my wardrobe has evolved. I find that I value my purchases more when I have researched them and know that they have good credentials and now I really wouldn’t get the same gratification that I used to from buying from a high street brand. In a nutshell, cherish what you have, mend things, alter things. When you decide to get rid of something think about how you are going to dispose of it.
Hair by Luke Benson at Frank using GHD Professional and Oribe
Makeup by Dina Catchpole at Frank Agency
Find out more about Antibad and to see the collection:
www.antibadstore.com |
Q:
¿Por qué no me deja validar que las variables no permitan valores negativos o ceros?
Mi problema es que el ejercicio me pide: validar que las variables no permitan
valores negativos o ceros; esto lo debe hacer en los métodos accesores o propiedades.
Y ese es el problema por que no lo hace y quiero que me muestre el mensaje que me diga no se puede tener un valor negativo
namespace Ejercicio8
{
class Trapecio
{
public double area = 0, bm = 0, bme = 0, h = 0;
public double Calculo()
{
area = (bm + bme) * h / 2;
return area;
}
public double Area
{
set
{
if (value >= 0)
{
area = value;
}
else
{
Console.WriteLine("No se puede tener un valor negativo.");
}
}
get
{
return area;
}
}
public double Bm
{
get
{
return bm;
}
set
{
if (bm >= 0)
{
value = bm;
}
else
{
Console.WriteLine("error"+ Bm);
}
}
}
public double Bme
{
get
{
return bme;
}
set
{
if (bme >= 0)
{
value = bme;
}
else
{
Console.WriteLine("error"+ Bme);
}
}
}
public double H
{
get
{
return h;
}
set
{
if (h >= 0)
{
value = h;
}
else
{
Console.WriteLine("error"+H);
}
}
}
}
class Program
{
static void Main(string[] args)
{
Trapecio t = new Trapecio();
Console.WriteLine("===Encontrar el area de un trapecio====");
Console.WriteLine("Digite la base mayor del trapecio:");
t.bm = Convert.ToInt32(Console.ReadLine());
Console.WriteLine("Ingrese la base menor del trapecio:");
t.bme = Convert.ToInt32(Console.ReadLine());
Console.Write("Digite la altura del trapecio: ");
t.h = Convert.ToInt32(Console.ReadLine());
t.Calculo();
Console.WriteLine("====== RESULTADOS =======");
Console.WriteLine("La base mayor ingresada es de:" + t.Bm);
Console.WriteLine("La base menor ingresada es de:" + t.Bme);
Console.WriteLine("La altura ingresada es de: " + t.H);
Console.WriteLine("El area del trapecio es de:" + t.Area);
Console.ReadKey();
}
}
}
}
A:
Las respuestas de LPZadkiel y Martin son complementarias. La solución es invertir la validación en la declaración del setter y getter, y cambiar el llamado de la variable en el método main. Así:
namespace Ejercicio8
{
class Trapecio
{
// Cambio de public por private
private double area = 0, bm = 0, bme = 0, h = 0;
public double Calculo()
{
area = (bm + bme) * h / 2;
return area;
}
public double Area
{
set
{
if (value >= 0)
{
area = value;
}
else
{
Console.WriteLine("No se puede tener un valor negativo.");
}
}
get
{
return area;
}
}
public double Bm
{
get
{
return bm;
}
set
{
// Cambio de variable a validar
if (value >= 0)
{
bm = value;
}
else
{
Console.WriteLine("error"+ Bm);
}
}
}
public double Bme
{
get
{
return bme;
}
set
{
// Cambio de variable a validar
if (value >= 0)
{
bme = value;
}
else
{
Console.WriteLine("error"+ Bme);
}
}
}
public double H
{
get
{
return h;
}
set
{
// Cambio de variable a validar
if (value >= 0)
{
h = value;
}
else
{
Console.WriteLine("error"+H);
}
}
}
}
class Program
{
static void Main(string[] args)
{
Trapecio t = new Trapecio();
Console.WriteLine("===Encontrar el area de un trapecio====");
Console.WriteLine("Digite la base mayor del trapecio:");
// Cambio de variable de llamado
t.Bm = Convert.ToInt32(Console.ReadLine());
Console.WriteLine("Ingrese la base menor del trapecio:");
// Cambio de variable de llamado
t.Bme = Convert.ToInt32(Console.ReadLine());
Console.Write("Digite la altura del trapecio: ");
// Cambio de variable de llamado
t.H = Convert.ToInt32(Console.ReadLine());
t.Calculo();
Console.WriteLine("====== RESULTADOS =======");
Console.WriteLine("La base mayor ingresada es de:" + t.Bm);
Console.WriteLine("La base menor ingresada es de:" + t.Bme);
Console.WriteLine("La altura ingresada es de: " + t.H);
Console.WriteLine("El area del trapecio es de:" + t.Area);
Console.ReadKey();
}
}
}
|
Media playback is unsupported on your device Media caption Eyewitness Ozy Licano: ''He started shooting and I was looking at his face''
A shooting at a family planning clinic in Colorado Springs has left two civilians and a police officer dead, with the suspected gunman under arrest.
Nine other people were injured during the standoff at the Planned Parenthood clinic, which lasted five hours before the suspect surrendered.
A number of people were trapped inside the building as shots were exchanged.
The motive remains unclear. The Planned Parenthood group has drawn anti-abortion protests in the past.
Colorado Springs Police Department identified the suspect as Robert Lewis Dear, from North Carolina.
"I want to convey to the loved ones of the victims, this is a terrible, terrible tragedy that occurred here in Colorado Springs today," Mayor John Suthers told a news conference.
"Obviously, we lost two civilian victims. We mourn the loss of a very brave police officer."
The dead policeman was named as Garrett Swasey, 44, who was married with two children.
President Barack Obama said: "We have to do something about the easy accessibility of weapons of war on our streets to people who have no business wielding them. Enough is enough."
He has previously expressed his frustration after other fatal shootings about not being able to do more on gun control.
Image copyright AFP Image caption The suspect has been named as Robert L Dear, 57
Image copyright Reuters Image caption Garrett Swasey, who was killed, was a campus police officer for the University of Colorado at Colorado Springs (UCCS)
Image copyright Reuters Image caption A man could be seen being taken into custody outside the centre
Colorado Springs Police Chief Peter Carey said five police officers were among the injured, who were being treated in local hospitals.
Police had sealed off streets around the centre as officers tried to make contact with the suspect.
"We did get officers inside the building," police Lt Catherine Buckley said.
What is Planned Parenthood?
A healthcare non-profit-making group with 59 affiliates and 700 clinics around the US
The largest single provider of abortion in the US
Its clinics provide many other healthcare services including cancer screening
Dates back to 1916 when social activist and nurse Margaret Sanger opened the first birth control and family planning centre in Brooklyn, New York
In the 1960s and 1970s, Planned Parenthood affiliates were at the fore of many court fights to make abortion legal
A controversial service
"They were able to shout to the suspect and make communication with him and at that point they were able to get him to surrender and he was taken into custody."
Police said the suspect had been seen carrying "some bags" into the building and teams were combing the area for possible explosives.
The suspect's car which was found nearby was also being checked for explosives, CNN reported.
Image copyright Getty Images Image caption Armed police patrolled the streets outside the centre during the stand-off
Image copyright Getty Images Image caption People who escaped the centre were led to safety by police
The manager of a nearby hair salon, Denise Speller, said she had heard as many as 20 gunshots in under five minutes.
She told a local newspaper she had seen one of two police officers appear to fall to the ground and the other attempt to get the wounded officer behind their police vehicle.
Police had told shoppers at a nearby centre to stay indoors.
In a statement, Planned Parenthood said it was not yet clear "if Planned Parenthood was in fact the target of this attack".
"Our concern is for the safety of our patients, staff and law enforcement," said its CEO, Vicki Cowart.
Planned Parenthood has been the focus of protests recently after an anti-abortion organisation secretly recorded a Planned Parenthood official discussing how to obtain aborted foetal tissue for medical research.
Pro-life advocates say this proves Planned Parenthood is selling foetal parts for profit - which is illegal - but this is disputed by the organisation.
New York's city police said it had deployed critical response vehicles to Planned Parenthood sites in the city because of the Colorado incident, but said there was no threat in the city. |
Low Cost Dumpster Rentals
3 Yard Dumpster in Eckert, CO
Free Price Quote, Call Now!
Need a Dumpster Rental in Eckert?
We have the commercial and residential dumpster service you need in Eckert.
At WasteRatings, we strive to bring you the best service at the best price available. We have a wide selection of dumpster sizes in stock in Eckert, CO, with flexible & timley pickup and delivery. Give us a call today for all your dumpster rental and roll off needs.
Give us a call today for a free dumpster quote. (970) 289-0677
More About Us
WasteRatings has made it super fast and easy to rent a dumpster. You only need 5 minutes!
When you are renting a 5, 10, 15, 20, 30 or 40 yard dumpster, you want a company you can trust with prices that make you smile. Give us a call today and see the difference we can make in your next construction or clean out project.
Free Consultation
Simply give us a call and we will help you figure out your dumpster rental needs.
Prompt Delivery
Our dumpsters usually go out same-day or next-day depending on when you call.
Competitive Pricing
We provide top-notch service, while going easy on your bottom line. What more could you ask for?
Hassle-Free Ordering
Our trained operators are here to give you a fast and hassle-free experience from start to finish.
Fast Pickup
Once you are done, simply give us a call and we will get your dumpster picked up right away!
Responsive Staff
Our staff is here to help whenever you need it. Simply pick up the phone and give us a call.
Get a Dumpster Quote Now!
Operators are Standing By
Our Dumpster Specialists are waiting by the phone to help you get started. However, there is some important information we need before we can give you a price quote and delivery time. Some things we need to know:
What size dumpster you will need
What you plan on putting in your dumpster
When you need us to drop your roll off container off
How long you plan on keeping your roll-off
City & zip code where you plan on placing your dumpster
Ready to get started?
Dumpsters are available* in Eckert.
Simply give us a call now at (970) 289-0677 to get started! We will help you pick a dumpster for your needs, schedule your delivery and take payment… all in less than 10 minutes!
Zip Codes in Service Area
Free Price Quote, Call Now!
Dumpster Rental Info Center
City Regulations on Dumpster Rental
Most cities or municipalities do not have many regulations regarding dumpster rental as long as you keep the dumpster entirely on your own property during the rental period. If you need to place your container on the street in any way, you will likely have to obtain a permit from the appropriate building permit office in your town. You can check with your local public works department to be sure.
Most Eckert dumpster rental companies will take care of securing this permit for you if you rent from them. Make sure that if you’re planning to put the dumpster on the street, the company has made the appropriate arrangements. You should also ensure that you get the permit in a timely manner and at the correct cost. If you thought the dumpster company was getting a permit and they did not, you will be the one who will have to pay the fine that is issued by the authorities.
Rent a Dumpster As soon as Updating Your Farmhouse in Eckert, CO
Speedy service for instance next day 3 Yard Dumpster in Eckert, CO is yet another fantastic object to look for. For anyone who is in a hurry to obtain the project commenced and accomplished, this is a should have option.
Work out which size hired dumpster you will need. The 3 Yard Dumpster in Eckert, CO company must have at least several sizes-small, medium and larger.
Nearly any kind of business will be needing some sort of 3 Yard Dumpster in Eckert, CO with regards to property. With respect to the type of business will certainly reflect which service is needed, and will also reveal what size dumpster has to be rented. Some of the available pick up options ranges from every day, to each week, to actually monthly pick-ups. Specialty dumpsters pertaining to recyclable goods are also available for you to rent out to organizations. If you need a compacting dumpster pertaining to boxes and cardboard solutions, this too can be rented through the same business that rents out any other type of dumpster.
A lot of the 3 Yard Dumpster in Eckert, CO companies cost fee based the time period of the rental and tonnage of the particles disposed to the dumpster. Be quite specific about the time period of your rental before you place an order. Though many of the companies deal with instant jobs, it is always best to schedule your delivery and pick up in advance so that you have ample time to make out room for dumpster when they get it to your area.
This is where any 3 Yard Dumpster in Eckert, CO service stomach in useful. The company not simply provides the dumpster, in addition, they discard your waste in your case. So, this may be one a smaller amount thing you would probably have to worry about, specifically after a wedding ceremony.
All objects can be dumped in one dumpster. They don't need to be divided for example: wine glass, metal, solid wood and sheetrock can all be disposed of from the same dumpster. Neighborhood trash pick up regulations reduce the amount of waste that is available pertaining to pickup from the city. They also require the getting of all waste and the removing of any recyclables. Using a construction 3 Yard Dumpster in Eckert, CO creates trash fingertips much easier and quicker mainly because it does not require your bagging associated with trash or even the separating associated with recyclables.
It is time for every individual to critically understand the need to protect the planet. In spite of the govt taking a lot of measures guard the environment; the planet is still impacted due to the smog that is created in the country. It's very important for the govt and individuals to work in concurrent to protect the planet.
If you want your dumpster placed on the driveway, you might lay down numerous pieces of particle board to protect the location from any damage.
Tired of every one of the clutter taking up valuable room in your home as well as on your lawn? Next don't loose time waiting for spring for you to swing in to cleaning function. With simple and inexpensive 3 Yard Dumpster in Eckert, COs from a local fingertips company, you can clear away years of accumulated useless in a single day and reclaim the real estate pertaining to better work with.
Once the property has been satisfied of the tenant's points, it's time to target repairs. Evictees will often be terrible tenants (they had for being evicted, in the end!), and they are capable of doing some rather extensive injury to your property that must definitely be repaired before you rent it out again. Odds are that you will find to replace your carpeting from the rental; few renters leave in one piece carpeting. Should the former lodger had dogs and cats, it is essential to replace the carpet to remove all contaminants in the air. However, if perhaps there has been virtually any damage done to the wall space or painting is necessary, save your time and difficulty by keeping people old carpets and rugs installed to operate as drop cloths while you repair and repaint. In the event the repairs are carried out, rip available and replace the old carpeting. Perform virtually any needed preservation on kitchen appliances and lighting fixtures, and then start showing the property for being rented again. Better chance with your future tenants!
Know the Difference Between Full-Service and Limited Service-Dumpster Rentals Prior to deciding to Hire in Eckert
In the year you may want to start off school makeovers that need to occur now. As time goes by, schools are obtaining more and more packed. Additions are necessary as well as class room sizes manufactured larger. For the duration of spring break you could have a contractor lined up to start the particular renovations. To acquire the makeovers done rapidly you will need to find a person that housing costs next day dumpsters. The actual contractors may easily do their work and also clean up soon after with a dumpster suitable next to them. Make sure you cope with a company that has dumpsters with wheels on the bottom. The actual contractors will be able to easily move the dumpster from one area to another even though working.
Controlling these big tons of junk on personal would be a pain and requires lots of investment in regards to time and money. Ones painstaking analysis to find an inexpensive way of removing the trash inside shortest period possible would certainly end up in 3 Yard Dumpster in Eckert services.
Dumpsters are usually accessible in different the particular sizes that range between 10 property and 40 yard. You have to be extra careful while picking out the size of dumpster mainly because choosing the completely wrong sized dumpster in addition to required will impose further charges as you have to go for multiple pickups to be able to dispose your current trash if the dumpster is tiny in size. You can be paying for the free space if you choose a dumpster that's large wide.
A major cleanout can be stressful and also exhausting. No matter if you're planning to have an upcoming cross-country move into a smaller household, cleaning out a senior relative's home just before an property sale or even foreclosure, or even gut-renovating your office and also installing all-new lighting fixtures, all of the junk you throw has to head out somewhere. It is possible to divide the garbage and junk up in small plenty and get back and forth to your dump and also the recycling core, wasting time, fuel, and also. Or you can may well avoid the hassle by means of hiring a specialist junk transporting team to complete the hard work and disposal for you. Just what does a junk hauler do? The treatment depends on the specific company, thus consider your requirements your budget in order to making a hiring decision.
Whenever you embark on a major cleaning, development, or renovation project, the very last thing you want to stress about is getting lessen the ending debris. Working on the project is hard plenty of. There's no good reason that you should be weighted down with luggage and piles of junk afterward. It may clutter the property, which often directly negates the aim of doing the work initially and can warrant special excursions to the remove or other disposal facility after the fact. Instead of fixing your junk removal inside a piecemeal way, it usually is worth your while to review renting a dumpster. While using a dumpster in your yard may seem like a pace backwards, this task simplifies the particular cleanup process. You can rent a dumpster for as long as you will need, you can put all of your trash and also debris inside it, and then you could possibly get rid of everything in one dropped swoop when you're done. For anyone who is doing a home clean-out, construction cleanup, or other diy or renovation projects, renting a dumpster can be a excellent option.
Do You Know Where to Place The Roll-Off Dumpster At the Job Place in Eckert, CO
On the other side, also be aware of companies that happen to be very large. Often those corporations have so many customers that all customer gets very cold service. For the kids, losing just one customer will not be a big deal. This is also true if the organization is so big that it must be nationwide.
A brand new pool table for you personally and your close friends to get jointly and use is an excellent strategy for your kept up to date basement. In case you have children which might be a little older they will enjoy the fun room observing movies jointly or playing pool. You should understand where young kids are and they also are safe within your basement as an alternative to running around the area or away driving close to.
While the twenty garden dumpster will take care of entire home renovations and enormous basement as well as attic clear outs, massive demolitions might need something greater. A 25 yard dumpster was created to hold up to a few tons of throw away and particles and particles and will effortlessly fill in any time other sized dumpsters just will not do. When you are aware enough of your project and also the size of your rental waste as well as debris dumpsters you need to work with, you can actually get the dumpster you'll need for maximum results on your challenge.
The family bedroom is an important position where you and your loved ones hang out. Movies and popcorn are distributed, conversations, boardgames or Wii system games, etc. In the evening everyone congregates to the family area so it is effortlessly understood the reason one would need a very at ease and lovely family area.
Basements certainly are a huge challenge that will require a 3 Yard Dumpster in Eckert, CO. You may be losing old waste that you have obtained over the past forty years or you may want to remodel ones basement. In any event, a overnight 3 Yard Dumpster in Eckert, CO is important to get the job done.
The trash accumulated on the business spot will make your business famous for the garbage piles for your location instead of product bought. you are required to do necessary home work to choose the best vendor at your organization location. A sensible investment often gives you the top return.
Therefore forget the price hikes as well as disposal limits of visiting your regional remove. Now you can clear away all of your household garbage as well as debris a single easy move without previously leaving home. Get in touch with your local how you can schedule ones convenient 3 Yard Dumpster in Eckert, CO, before spring.
Next you can place your fresh stained baseboards onto the reduce wall as well as floor. When you ordered brand-new doors therefore you or the door company have got stained them, now is the time to get them on the hinges.
Minor and major businesses work with dumpsters as well. Regardless of whether a business is definitely moving or perhaps the type of organization that has a lot of waste, it's a good idea to help them to have a dumpster that they may clean up along with at the end of the time.
Deciding on the suitable size dumpster will be the first step. Almost all refuse goods or supplies should be positioned safely as well as completely inside dumpster and there should be nothing herniated. Locating the appropriate dumpster for your needs as well as hauling away your garbage can be established based on the challenge. Smaller initiatives using entail cleaning out cellars or car ports while greater projects incorporate renovations as well as kitchen refurbishments. Small dumpsters can be purchased that can be wheeled about the home and choice and large dumpsters are for sale to projects that involve construction as well as renovation supplies.
Request a Call Back!
Still need a quote but don't want to talk on the phone right now? Fill out the form below and we will get back to you!
Please use the form below to contact us. We look forward to working with you and will get back to you as soon as possible. All fields are required |
A protest against the sale of lads’ mags was staged at numerous Tesco stores around the country today.
In an effort to force Britain’s largest retailer to ditch magazines such as Zoo and Nuts, the Lose the Lads’ Mags campaign pitched up in cities ranging from Manchester, London, Portsmouth and Glasgow from midday for one hour.
The campaign, spearheaded by women’s rights groups Object and UK Feminista, is against the sexualisation and objectification of women, as well as the exposure of such images to children.
The group says that lads’ mags ‘fuel sexist behaviours and attitudes underpinning violence against women’, adding: ‘Tesco would never allow “girlie calendars” on their office walls. Why are they choosing to stock degrading, pornographic lads’ mags on their shelves?’
Activists said it would talk to consumers outside Tesco stores and collect signatures in a ‘family-friendly protest’.
Lads’ mags are currently sold in Co-operative stores, although the store has asked them to cover up in modesty bags (Picture: PA)
The latest wave of action comes weeks after the Co-operative issued an ultimatum to numerous lads’ mags such as Zoo and Front, which gave them the option of concealing sexy images in modesty bags before September 9 – or expect to be thrown out of stores.
Kat Banyard, the founder of UK Feminista, said: ‘As long as Tesco sells lads’ mags like Nuts and Zoo its claims of being a responsible corporate citizen are a farce.
‘Tesco needs to put women’s safety before profit and lose the lads’ mags.’
The row over the depiction of women and sexualisation of young girls has intensified over the last few weeks, with David Cameron promising a wave of reforms that would restrict the exposure of pornographic material to children, particularly on the internet.
He did, however, stop short of condemning The Sun’s infamous Page 3, saying in July: ‘This is an area where we should leave it to consumers to decide, rather than to regulators.’
Campaigners are trying to abolish the sale of lads’ mags from supermarkets and other retailers (Picture: PA Wire/Press Association Images)
The No More Page 3 campaign, which began last year, has welcomed today’s protests and said that they, too, were not against voicing their concerns in a day of action outside of retailers.
‘Protests are a great way of engaging with the public about the issue,’ said founder Lucy Holmes.
‘We support the Lads Mags campaign because since we started the No More Page 3 petition last year, we’ve lost count of the amount of parents who have written to us saying “can you do something about lads’ mags and the cover of The Sport and Star as well.”‘ |
namespace Graphviz4Net.WPF.Example
{
using System;
using System.Windows;
using System.Windows.Controls.Primitives;
using System.Windows.Input;
using System.Threading;
using System.Windows.Controls;
using System.Diagnostics;
using System.Reflection;
using System.IO;
using System.IO.Compression;
using System.Windows.Media.Imaging;
using System.Threading.Tasks;
using RichTextBlockSample.HtmlConverter;
using System.Windows.Markup;
using System.Windows.Documents;
using System.Collections.Generic;
using System.Linq;
using Microsoft.Win32;
using System.Windows.Media.Animation;
using System.Management.Automation;
using System.Collections.ObjectModel;
using System.Windows.Media;
using System.DirectoryServices;
//using System.Drawing;
//using System.Windows.Forms;
public partial class MainWindow : Window
{
private MainWindowViewModel viewModel;
public bool showLegend = false, firstTimeShowACL = false;
public int whatToRun;
public bool quitApplication = false;
public string logFileName = "../../ZBANG/log.txt";
string totalString = "";
public MainWindowViewModel ViewModel { get; set; }
public enum TABITEMS
{
ACLLIGHT,
SKELETONKEY,
SIDHISTORY,
RISKYSPNS,
MYSTIQUE,
EASYPEASY
}
private static BitmapImage GetImage(string imageUri)
{
var bitmapImage = new BitmapImage();
bitmapImage.BeginInit();
bitmapImage.UriSource = new Uri( "pack://siteoforigin:,,,/" + imageUri, UriKind.RelativeOrAbsolute );
bitmapImage.EndInit();
return bitmapImage;
}
public MainWindow()
{
try
{
this.viewModel = new MainWindowViewModel();
ViewModel = viewModel;
this.DataContext = viewModel;
this.showLegend = true;
InitializeComponent();
//MessageBox.Show("testing456");
using( PowerShell PowerShellInstance = PowerShell.Create() )
{
// use "AddScript" to add the contents of a script file to the end of the execution pipeline.
// use "AddCommand" to add individual commands/cmdlets to the end of the execution pipeline.
PowerShellInstance.AddScript( "$PSVersionTable" );
Collection<PSObject> PSOutput = PowerShellInstance.Invoke();
string outputlog = "-------------------------------------------\n";
foreach( PSObject outputItem in PSOutput )
{
// we should be getting a hashtable... dump it to the log file
if( outputItem != null )
{
//TODO: do something with the output item
System.Collections.Hashtable hash = outputItem.BaseObject as System.Collections.Hashtable;
if( ((System.Version)hash["PSVersion"]).Major < 3)
MessageBox.Show("The zBang Tool needs PowerShell version >= 3.0\nPlease download an updated version from\nhttps://www.microsoft.com/en-us/download/details.aspx?id=34595");
foreach( string key in hash.Keys )
{
if( hash[key].GetType() == typeof( System.Version[] ) )
{
System.Version[] vers = (System.Version[])hash[key];
string verstr = "";
foreach( System.Version ver in vers )
{
verstr += ver.ToString() + ",";
}
string format = "{0,-25}\t{1,-10}";
string result = string.Format( format, key, verstr );
outputlog += /*key + "\t\t\t" + verstr +*/result + "\n";
}
else
{
string format = "{0,-25}\t{1,-10}";
string result = string.Format( format, key, hash[key] );
outputlog += result + "\n";
}
}
}
}
outputlog += "-------------------------------------------\n";
File.AppendAllText( /*"..\\..\\ZBANG\\log.txt"*/logFileName, "\n\nzBang Launched at " + DateTime.Now.ToString() + "\n" + outputlog + "\n" );
} // endusing powershell
if( !GetDotNetVersion.Get45PlusFromRegistry( logFileName ) )
{
MessageBox.Show(".NET Version is below 4.5\nPlease download .NET 4.5 from microsoft.com...\nPlease download from https://www.microsoft.com/en-us/download/details.aspx?id=30653 \nAborting...", "Info", MessageBoxButton.OK, MessageBoxImage.Hand );
System.Windows.Application.Current.Shutdown();
return;
}
//
// report to temp log file all the domain names in the forest
try
{
List<string> domainNames;
int count = viewModel.enumerateDomainInForest( out domainNames );
string _outputlog = "\n" + count.ToString() + " domain(s) in forest:\n";
int c = 0;
foreach( string dom in domainNames )
{
c++;
_outputlog += "(" + c.ToString() + ") " + dom + "\n";
}
File.AppendAllText( logFileName, _outputlog );
}
catch
{
}
/** LOAD IMAGES **/
//string ImagesPath = "pack://application:,,/Graphviz4Net.WPF.Example;component/Images/aetosdios_trans.png";
string ImagesPath = "./Images/flashlight.png";
Uri uri = new Uri( ImagesPath, UriKind.RelativeOrAbsolute );
BitmapImage bitmap = new BitmapImage( uri );
imgACLight.Source = bitmap;
imgSkeleton.Source = new BitmapImage( new Uri( @"Images/aetosdios_trans.png", UriKind.Relative ) );
imgRisky.Source = new BitmapImage( new Uri( @"./Images/aetosdios_trans.png", UriKind.Relative ) );
imgSID.Source = new BitmapImage( new Uri( @"./Images/aetosdios_trans.png", UriKind.Relative ) );
imgMistique.Source = new BitmapImage( new Uri( @"./Images/aetosdios_trans.png", UriKind.Relative ) );
bulletACL.Source = new BitmapImage( new Uri( @"./Images/flashlight.png", UriKind.Relative ) );
bulletSkeleton.Source = new BitmapImage( new Uri( @"./Images/key.png", UriKind.Relative ) );
bulletSID.Source = new BitmapImage( new Uri( @"./Images/theater.png", UriKind.Relative ) );
bulletSPN.Source = new BitmapImage( new Uri( @"./Images/clerk.png", UriKind.Relative ) );//Source="pack://application:,,,/Images/bullet_grey.png"
bulletMystique.Source = new BitmapImage( new Uri( @"./Images/role-playing-game.png", UriKind.Relative ) );
imgArrowLeft.Source = new BitmapImage( new Uri( @"./Images/arrow_left.png", UriKind.Relative ) );
/* NS GITHUB
imageLogo.Source = new BitmapImage(new Uri(@"./Images/LogoSeparator_b.png", UriKind.Relative));
*/
ToolTipService.ShowDurationProperty.OverrideMetadata( typeof( DependencyObject ), new FrameworkPropertyMetadata( Int32.MaxValue ) );
tabControl.Visibility = Visibility.Hidden;
GraphLayout.Visibility = Visibility.Hidden;
this.showmach.IsChecked = true;
this.showmach2.IsChecked = true;
/*
this.AddNewEdge.Click += AddNewEdgeClick;
this.AddNewPerson.Click += AddNewPersonClick;
this.UpdatePerson.Click += UpdatePersonClick;
*/
zoomControl.MaxZoom = 3.0;
zoomControl.ZoomDeltaMultiplier = 20.0;
this.WindowStartupLocation = WindowStartupLocation.CenterScreen;
this.WindowState = WindowState.Maximized;
System.Reflection.Assembly assembly = System.Reflection.Assembly.GetExecutingAssembly();
FileVersionInfo fvi = FileVersionInfo.GetVersionInfo( assembly.Location );
string version = fvi.FileVersion;
//this.Title = "Zbang Tool Suite, ver." + version + " CyberArk 2018";
this.lblCursorPosition.Text = "zBang Tool Suite, ver." + version /*+ " CyberArk 2018" GITHUB */;
// set background image for the grid
string _ImagesPath = "pack://application:,,/Graphviz4Net.WPF.Example;component/Images/logo_color.png";
ImageBrush myBrush = new ImageBrush();
/* NS GITHUB
Image image = new Image();
image.Source = new BitmapImage( new Uri( _ImagesPath, UriKind.RelativeOrAbsolute ) );
myBrush.ImageSource = image.Source;
zoomControl.Background = myBrush;
zoomControl.Background.Opacity = 0.09;
*/
}
catch ( Exception e )
{
System.Windows.MessageBox.Show( e.Message );
}
} // end MainWindow
#if zero
using System;
using System.IO;
public class BasicTest
{
// try get images from DC
static string GetUserPicture(string userName)
{
string LDAP = "LDAP://AETOSDIOS";
//using (DirectoryEntry dirEntry = new DirectoryEntry(/*LDAP, null, null, AuthenticationTypes.Secure)*/)
{
using (DirectorySearcher dsSearcher = new DirectorySearcher( /*dirEntry*/))
{
dsSearcher.Filter = "(&(objectCategory=person)(sAMAccountName=" + userName + "))";
SearchResult result = dsSearcher.FindOne();
using (DirectoryEntry user = new DirectoryEntry(result.Path))
{
byte[] data = user.Properties["thumbnailPhoto"].Value as byte[];
if (data != null)
{
using (MemoryStream s = new MemoryStream(data))
{
byte[] byteArr = s.ToArray();
string binStr = Convert.ToBase64String(byteArr);
return binStr;
}
}
return null;
}
}
}
} // endfunc
}
#endif
void runLaunchWindow()
{
selectTogglesForms atoggle = new selectTogglesForms();
bool dresult = (bool)atoggle.ShowDialog();
if( !dresult )
{
//System.Windows.Application.Current.Shutdown();
return;
}
toggleACLight.IsChecked = false;
toggleSkeleton.IsChecked = false;
toggleSIDHistory.IsChecked = false;
toggleRisky.IsChecked = false;
toggleMystique.IsChecked = false;
int aq = atoggle.whoIsSelected & (1 << (int)MainWindow.TABITEMS.ACLLIGHT);
if( aq != 0 )
toggleACLight.IsChecked = true;
aq = atoggle.whoIsSelected & (1 << (int)MainWindow.TABITEMS.SKELETONKEY);
if( aq != 0 )
toggleSkeleton.IsChecked = true;
aq = atoggle.whoIsSelected & (1 << (int)MainWindow.TABITEMS.SIDHISTORY);
if( aq != 0 )
toggleSIDHistory.IsChecked = true;
aq = atoggle.whoIsSelected & (1 << (int)MainWindow.TABITEMS.RISKYSPNS);
if( aq != 0 )
toggleRisky.IsChecked = true;
aq = atoggle.whoIsSelected & (1 << (int)MainWindow.TABITEMS.MYSTIQUE);
if( aq != 0 )
toggleMystique.IsChecked = true;
if( atoggle.isLaunched == 1 )
{
this.LaunchZbang.RaiseEvent( new RoutedEventArgs( ButtonBase.ClickEvent ) );
}
else if( atoggle.isLaunched == 0 )
this.Launch.RaiseEvent( new RoutedEventArgs( ButtonBase.ClickEvent ) );
else if( atoggle.isLaunched == 2)
{
this.ImportButton.RaiseEvent( new RoutedEventArgs( ButtonBase.ClickEvent ) );
}
} // endfunc launch window
void UpdatePersonClick(object sender, RoutedEventArgs e)
{
/*
this.viewModel.UpdatePersonName = (string) this.UpdatePersonName.SelectedItem;
this.viewModel.UpdatePerson();
*/
}
private void AddNewPersonClick(object sender, RoutedEventArgs e)
{
this.viewModel.CreatePerson();
}
private void AddNewEdgeClick(object sender, RoutedEventArgs e)
{
/*
this.viewModel.NewEdgeStart = (string) this.NewEdgeStart.SelectedItem;
this.viewModel.NewEdgeEnd = (string)this.NewEdgeEnd.SelectedItem;
this.viewModel.CreateEdge();
*/
}
int pressed = 0;
private void ToggleButton_Checked(object sender, RoutedEventArgs e)
{
rtbACL.Visibility = Visibility.Hidden; // hide the rich text ACL view
pressed++;
Launch.IsEnabled = true;
LaunchZbang.IsEnabled = true;
/*
ToggleButton tb = (ToggleButton)sender;
tb.Background = System.Windows.Media.Brushes.LightGreen;
tb.UpdateLayout();
*/
}
private void ToggleButton_UnChecked(object sender, RoutedEventArgs e)
{
rtbACL.Visibility = Visibility.Hidden; // hide the rich text ACL view
pressed--;
if( pressed <= 0 )
{
Launch.IsEnabled = false;
LaunchZbang.IsEnabled = false;
}
/*
ToggleButton tb = (ToggleButton)sender;
tb.Background = System.Windows.Media.Brushes.LightGreen;
tb.UpdateLayout();
*/
}
private void launchButton_clicked(object sender, RoutedEventArgs e)
{
rtbACL.Visibility = Visibility.Hidden; // hide the rich text ACL view
zoomControl.MaxZoom = 3.0;
zoomControl.Mode = WPFExtensions.Controls.ZoomControlModes.Fill;
tabControl.Visibility = Visibility.Visible;
GraphLayout.Visibility = Visibility.Visible;
//---> disable all the tabitems not in the launch game
if( !(bool)toggleACLight.IsChecked )
ACLLight.IsEnabled = false;
else
ACLLight.IsEnabled = true;
if( !(bool)toggleSkeleton.IsChecked )
SkeletonItem.IsEnabled = false;
else
SkeletonItem.IsEnabled = true;
if( !(bool)toggleSIDHistory.IsChecked )
SIDHistoryItem.IsEnabled = false;
else
SIDHistoryItem.IsEnabled = true;
if( !(bool)toggleRisky.IsChecked )
SPNItem.IsEnabled = false;
else
SPNItem.IsEnabled = true;
if( !(bool)toggleMystique.IsChecked )
MystiqueItem.IsEnabled = false;
else
MystiqueItem.IsEnabled = true;
/*
if (!(bool)toggleEasy.IsChecked)
EasyPeasy.IsEnabled = false;
else
EasyPeasy.IsEnabled = true;
*/
// if acl light button pressed?
if( (bool)toggleACLight.IsChecked )
{
if( !ACLLight.IsSelected)
firstTimeShowACL = true;
ACLLight.IsSelected = true;
if( viewModel.scanACLoutputForDomains( null, "ACLight Discovered Domains" ) == 1 )
this.viewModel.reformatCardsGraph( null, true );
else
viewModel.onScreen = MainWindowViewModel.ON_SCREEN.DOMAIN_CARDS;
}
else if( (bool)toggleSkeleton.IsChecked )
{
SkeletonItem.IsSelected = true;
//this.viewModel.showSkeletonKeyResults(true);
}
else if( (bool)toggleSIDHistory.IsChecked )
{
SIDHistoryItem.IsSelected = true;
}
else if( (bool)toggleRisky.IsChecked )
{
SPNItem.IsSelected = true;
}
else if ((bool)this.toggleMystique.IsChecked)
{
MystiqueItem.IsSelected = true;
}
/*
else if ((bool)toggleEasy.IsChecked)
{ }
*/
} // endfunc
private void menuShowMachines(object sender, RoutedEventArgs e)
{
if( sender == showmach || sender == showmach2 )
{
showmach.IsChecked = !showmach.IsChecked;
showmach2.IsChecked = !showmach2.IsChecked;
showLegend = showmach.IsChecked;
//this.viewModel.reformatCardsGraph();
}
}
private void menuRefresh(object sender, RoutedEventArgs e)
{
}
private void tabItemACL_Clicked(object sender, MouseButtonEventArgs e)
{
}
private void tabItemSkeleton_Clicked(object sender, MouseButtonEventArgs e)
{
}
private void tabItemSID_Clicked(object sender, MouseButtonEventArgs e)
{
}
private void tabItemSPN_Clicked(object sender, MouseButtonEventArgs e)
{
}
private void tabItemEasy_Clicked(object sender, MouseButtonEventArgs e)
{
}
private void tabItemMystique_Clicked(object sender, MouseButtonEventArgs e)
{
}
private void TabControl_SelectionChanged(object sender, System.Windows.Controls.SelectionChangedEventArgs e)
{
rtbACL.Visibility = Visibility.Hidden; // hide the rich text ACL view
zoomControl.MaxZoom = 3.0;
zoomControl.Mode = WPFExtensions.Controls.ZoomControlModes.Fill;
int sel = tabControl.SelectedIndex;
if( sel == (int)TABITEMS.ACLLIGHT && Launch.IsEnabled && !firstTimeShowACL)
{
if (viewModel.scanACLoutputForDomains(null, "ACLight Discovered Domains") == 1)
{
this.viewModel.reformatCardsGraph(null, true);
}
else
viewModel.onScreen = MainWindowViewModel.ON_SCREEN.DOMAIN_CARDS;
}
else if( sel == (int)TABITEMS.SKELETONKEY && Launch.IsEnabled )
{
this.viewModel.showSkeletonKeyResults( true );
}
else if( sel == (int)TABITEMS.MYSTIQUE && Launch.IsEnabled )
{
viewModel.onScreen = MainWindowViewModel.ON_SCREEN.RISKYSPN_ON_SCREEN;
Mystique.Run( null, this.viewModel.mysticInputFile );
}
else if( sel == (int)TABITEMS.SIDHISTORY && Launch.IsEnabled )
{
viewModel.onScreen = MainWindowViewModel.ON_SCREEN.SIDHISTORY_ON_SCREEN;
SIDHistory.Run(null, this.viewModel.sidHistoryInputFile);
}
else if( sel == (int)TABITEMS.RISKYSPNS && Launch.IsEnabled )
{
int ret;
if ((ret = viewModel.scanACLoutputForDomains(this.viewModel.RISKYSPNInputFile, "RiskySPNs Discovered Domains")) <= 1)
{
viewModel.onScreen = MainWindowViewModel.ON_SCREEN.RISKYSPN_ON_SCREEN;
riskySPNs.Run(null, this.viewModel.RISKYSPNInputFile);
}
else
viewModel.onScreen = MainWindowViewModel.ON_SCREEN.DOMAIN_CARDS_RISKY_SPN;
}
firstTimeShowACL = false;
} // endfunc
private void labelMouseDown(object sender, MouseButtonEventArgs e)
{
if( rtbACL.Visibility == Visibility.Visible)
{
rtbACL.Visibility = Visibility.Hidden;
return;
}
if( sender.GetType() == typeof( TextBlock ) )
{
TextBlock block = (TextBlock)sender;
string selectedText = block.Text;
if( this.ViewModel.onScreen != MainWindowViewModel.ON_SCREEN.ACLLIGHT_ON_SCREEN )
return;
string curDir = Directory.GetCurrentDirectory();
StreamReader sr = File.OpenText( String.Format( "{0}/../../ZBANG/ACLight Attack Path Update.html", curDir ) );
TextBox.Text = sr.ReadToEnd();
sr.Close();
/*
// select the required text
string richText = new TextRange( rtbACL.Document.ContentStart, rtbACL.Document.ContentEnd ).Text;
int iii = richText.IndexOf( selectedText );
TextPointer text = rtbACL.Document.ContentStart;
TextPointer startPos = text.GetPositionAtOffset( iii );
TextPointer endPos = text.GetPositionAtOffset( iii + 40 );
var textRange = new TextRange( startPos, endPos );
textRange.ApplyPropertyValue( TextElement.FontWeightProperty, FontWeights.Bold );
*/
rtbACL.Visibility = Visibility.Visible;
}
}
private Person highlightedPerson = null;
private DateTime showtime;
public int allowMouseLeave;
private void Border_MouseEnter(object sender, System.Windows.Input.MouseEventArgs e)
{
if( sender.GetType() == typeof(System.Windows.Shapes.Path))
{
System.Windows.Shapes.Path pth = (System.Windows.Shapes.Path)sender;
pth.Stroke = System.Windows.Media.Brushes.Gold;
return;
}
if( sender.GetType() == typeof( System.Windows.Controls.TextBlock) )
{
System.Windows.Controls.TextBlock pth = (System.Windows.Controls.TextBlock)sender;
pth.Background= System.Windows.Media.Brushes.Gold;
return;
}
Person ppp = (Person)(((System.Windows.Controls.Border)sender).DataContext);
if( ppp != highlightedPerson && this.viewModel.onScreen == MainWindowViewModel.ON_SCREEN.ACLLIGHT_ON_SCREEN )
{
this.viewModel.highlightEdges( ppp );
highlightedPerson = ppp;
showtime = DateTime.Now;
allowMouseLeave = 0;
}
else if( ppp == highlightedPerson && allowMouseLeave == 0 )
allowMouseLeave = 1;
}
private void OnGraphLayoutUpdated(object sender, EventArgs e)
{
/*
if ((DateTime.Now - showtime).TotalMilliseconds < 2000)
allowMouseLeave = 1;
*/
}
private void LabelClicked(object sender, System.Windows.Input.MouseEventArgs e)
{
}
private void Border_MouseLeave(object sender, System.Windows.Input.MouseEventArgs e)
{
if( sender.GetType() == typeof( System.Windows.Shapes.Path ) )
{
System.Windows.Shapes.Path pth = (System.Windows.Shapes.Path)sender;
pth.Stroke = System.Windows.Media.Brushes.Black;
return;
}
if( sender.GetType() == typeof( System.Windows.Controls.TextBlock ) )
{
System.Windows.Controls.TextBlock pth = (System.Windows.Controls.TextBlock)sender;
pth.Background = System.Windows.Media.Brushes.Transparent;
return;
}
if( allowMouseLeave != 1 )
return;
Person ppp = (Person)(((System.Windows.Controls.Border)sender).DataContext);
if(/*ppp == highlightedPerson &&*/ this.viewModel.onScreen == MainWindowViewModel.ON_SCREEN.ACLLIGHT_ON_SCREEN )
{
this.viewModel.dehighlightEdges( ppp );
highlightedPerson = null;
}
}
private void GoBackButton(object sender, RoutedEventArgs e)
{
rtbACL.Visibility = Visibility.Hidden; // hide the rich text ACL view
highlightedPerson = null;
allowMouseLeave = 0;
showtime = new DateTime( 0 );
if( this.viewModel.onScreen == MainWindowViewModel.ON_SCREEN.ACLLIGHT_ON_SCREEN )
this.viewModel.reformatCardsGraph( null, true );
else if( this.viewModel.onScreen == MainWindowViewModel.ON_SCREEN.CARDS_ON_SCREEN_WITH_DOMAIN_SELECTION )
{
this.viewModel.scanACLoutputForDomains( null, "ACLight Discovered Domains" );
viewModel.onScreen = MainWindowViewModel.ON_SCREEN.DOMAIN_CARDS;
}
else if( this.viewModel.onScreen == MainWindowViewModel.ON_SCREEN.RISKYSPN_WITH_DOMAIN_SELECTION )
{
if( viewModel.scanACLoutputForDomains( this.viewModel.RISKYSPNInputFile, "RiskySPN Discovered Domains" ) == 1 )
{
viewModel.onScreen = MainWindowViewModel.ON_SCREEN.RISKYSPN_ON_SCREEN;
riskySPNs.Run( null, this.viewModel.RISKYSPNInputFile );
backButton.Visibility = Visibility.Hidden;
}
else
{
viewModel.onScreen = MainWindowViewModel.ON_SCREEN.DOMAIN_CARDS_RISKY_SPN;
backButton.Visibility = Visibility.Hidden;
}
}
}
private void CheckBox_Checked(object sender, RoutedEventArgs e)
{
CheckBox a = (CheckBox)sender;
object der = (object)a.Parent;
if( typeof( Person ) == der.GetType() )
{
Person person = (Person)der;
person.BackColor = "LightGreen";
}
}
private void CheckBox_Unchecked(object sender, RoutedEventArgs e)
{
}
private void CheckBox_Indeterminate(object sender, RoutedEventArgs e)
{
}
/**
* @brief LAUNCH POWERSHELL MODULES
**/
ProgressWindow prg;
private void launchPowershellButton_clicked(object sender, RoutedEventArgs e)
{
rtbACL.Visibility = Visibility.Hidden; // hide the rich text ACL view
/*
. $PSScriptRoot\SourceFiles\RiskySPN-master\Get-PotentiallyCrackableAccounts.ps1
Report-PotentiallyCrackableAccounts -Type CSV -Path $PSScriptRoot\Results\RiskySPNs -DoNotOpen -Name RiskySPNs -test
*/
if( whatToRun == 0 )
{
prg = new ProgressWindow();
prg.Show();
prg.textBlockPowershell.Text += "\n";
if ( (bool)toggleSIDHistory.IsChecked )
whatToRun |= (1 << 2);
if( (bool)toggleSkeleton.IsChecked )
whatToRun |= (1 << 1);
if( (bool)toggleRisky.IsChecked )
whatToRun |= (1 << 3);
if( (bool)toggleMystique.IsChecked )
whatToRun |= (1 << 4);
if( (bool)this.toggleACLight.IsChecked )
whatToRun |= (1 << 0);
}
// END OF ROUND
else if( whatToRun == -1 )
{
prg.label1.Content = "*** FINISHED ***";
prg.progressBar1.Visibility = Visibility.Hidden;
// press the click on the launch button
Launch.RaiseEvent( new RoutedEventArgs( ButtonBase.ClickEvent ) );
whatToRun = 0;
//MessageBox.Show( logString, "Log Output for Launch at " + DateTime.Now.ToString(), MessageBoxButton.OK );
// NS 0.27 23/5/18
System.Reflection.Assembly assembly = System.Reflection.Assembly.GetExecutingAssembly();
FileVersionInfo fvi = FileVersionInfo.GetVersionInfo(assembly.Location);
string version = fvi.FileVersion;
string machineName = System.Environment.MachineName;
string username = System.Environment.UserName;
PerformanceCounter cpuCounter;
PerformanceCounter ramCounter;
cpuCounter = new PerformanceCounter(
"Processor",
"% Processor Time",
"_Total",
true
);
ramCounter = new PerformanceCounter("Memory", "Available MBytes", true);
File.AppendAllText( logFileName/*"..\\..\\ZBANG\\log.txt"*/, "Log Output for Launch at " + DateTime.Now.ToString() + ", version " + version + "\n" +
"Free RAM: " + ramCounter.NextValue().ToString()+"MB\n" +
"CPU Usage: " + cpuCounter.NextValue().ToString() + "%\n" +
"Computer Name: " + machineName + ", user Name: " + username + "\n" +
logString + "\n\n\n\n");
ExportButton.RaiseEvent( new RoutedEventArgs( ButtonBase.ClickEvent ) );
totalString = "";
prg.Close();
return;
}
if( (whatToRun & 1) != 0 ) // aclight
{
//string args = "/K \"powershell.exe -ExecutionPolicy Bypass -noprofile -command \"Import-Module './ACLight.psm1' -force ; Start-ACLsAnalysis\"\"";
string args = "-ExecutionPolicy Bypass -noprofile -command \"Import-Module './ACLight.ps1' -force ; Start-ACLsAnalysis\"";
// testing
////////// ???? windowRunPowerShell("Import-Module './../../ZBANG/ACLight-master/ACLight.psm1 ; Start-ACLsAnalysis'");
List<string> domainNames;
int count = viewModel.enumerateDomainInForest( out domainNames);
if( count > 1 )
{
domainSelection ds = new domainSelection( domainNames);
bool dialogresult = (bool)ds.ShowDialog();
//
if( ds.selection != -1)
{
args = args.Substring( 0, args.Length - 1) + " -domain '" + domainNames[ds.selection] + "'\"";
// NO NEED !!! MessageBox.Show( args, "error" );
}
}
#if zero
// start by scanning all comain in forest and if more than one domain is present pop up the question
int count = viewModel.enumerateDomainInForest();
if( count >= 1)
{
borderACLightQ.Visibility = Visibility.Visible;
DoubleAnimation myDoubleAnimation = new DoubleAnimation();
myDoubleAnimation.From = 0.0;
myDoubleAnimation.To = 1.0;
myDoubleAnimation.Duration = new Duration( TimeSpan.FromMilliseconds( 1000 ) );
// Configure the animation to target the button's Width property.
Storyboard.SetTargetName( myDoubleAnimation, borderACLightQ.Name );
Storyboard.SetTargetProperty( myDoubleAnimation, new PropertyPath( Border.OpacityProperty ) );
// Create a storyboard to contain the animation.
Storyboard myWidthAnimatedButtonStoryboard = new Storyboard();
myWidthAnimatedButtonStoryboard.Children.Add( myDoubleAnimation );
myWidthAnimatedButtonStoryboard.Begin( borderACLightQ );
return;
}
#endif
//prg.Show();
prg.progressBar1.Value = 50;
prg.label1.Content = "Now running ACLight test...";
logString = "";
logString += "\n--------------------------------------------\nPowershell - ACLight:\n--------------------------------------------\n\n";
totalString += "\n\n\n" + DateTime.Now.ToString() + "> Now Running ACLight\n";
File.AppendAllText( /*"..\\..\\ZBANG\\log.txt"*/logFileName, logString);
try
{
File.Delete("../../ZBANG/ACLight-master/Results/Privileged Accounts - Final Report.csv");
}
catch { }
var processStartInfo = new ProcessStartInfo
{
//FileName = @"../../ZBANG/ACLight-master/Execute-ACLight.bat",
FileName = "powershell.exe", //@"cmd.exe",
Arguments = args,
WorkingDirectory = "../../ZBANG/ACLight-master",
RedirectStandardOutput = true,
UseShellExecute = false,
CreateNoWindow = true
};
var process = Process.Start( processStartInfo );
process.OutputDataReceived += CaptureOutput;
process.ErrorDataReceived += CaptureOutput;
Thread runThread = new Thread( this.runThread );
runThread.Start( process );
} // endof aclight
else if( (whatToRun & (1 << 1)) != 0 ) // skeleton
{
//prg.Show();
prg.progressBar1.Value = 50;
prg.label1.Content = "Now running Skeleton Key test...";
totalString += "\n\n\n" + DateTime.Now.ToString() + "> Now Running SkeletonKey\n";
try
{
File.Delete("../../ZBANG/SkeletonKey_Scanner/Results/SkeletonKeyResults.csv");
}
catch { }
logString = "\n--------------------------------------------\nPowershell - SkeletonKey:\n--------------------------------------------\n\n";
File.AppendAllText( /*"..\\..\\ZBANG\\log.txt"*/logFileName, logString);
var processStartInfo = new ProcessStartInfo
{
FileName = @"Powershell.exe",
Arguments = "-ExecutionPolicy bypass -noprofile -command \"./SkeletonCheck.ps1\"",
WorkingDirectory = "../../ZBANG/SkeletonKey_Scanner",
RedirectStandardOutput = true,
UseShellExecute = false,
CreateNoWindow = true
};
var process = Process.Start( processStartInfo );
process.OutputDataReceived += CaptureOutput;
process.ErrorDataReceived += CaptureOutput;
Thread runThread = new Thread( this.runThread );
runThread.Start( process );
}
else if( (whatToRun & (1 << 2)) != 0 ) // SID History
{
//prg.Show();
prg.progressBar1.Value = 50;
prg.label1.Content = "Now running SID History test...";
totalString += "\n\n\n" + DateTime.Now.ToString() + "> Now Running SID History\n";
logString = "\n--------------------------------------------\nPowershell - SID History:\n--------------------------------------------\n\n";
File.AppendAllText( /*"..\\..\\ZBANG\\log.txt"*/logFileName, logString);
var processStartInfo = new ProcessStartInfo
{
FileName = @"Powershell.exe",
//Arguments = "-ExecutionPolicy Bypass -noprofile -Command \"& . ./SIDHistory_Scanner.ps1; Report-UsersWithSIDHistory '-Type \"CSV\"' '-Path \"Results\"' '-DoNotOpen'\"",
//Arguments = "-ExecutionPolicy Bypass -noprofile -Command \"& { ./SIDHistory_Scanner.ps1; Report-UsersWithSIDHistory -Type CSV -Path Results -DoNotOpen }\"",
Arguments = "-ExecutionPolicy Bypass -noprofile -Command \"& { ./SIDHistory_Scanner.ps1 }\"",
WorkingDirectory = "../../ZBANG/SIDHistory",
RedirectStandardOutput = true,
UseShellExecute = false,
CreateNoWindow = true
};
var process = Process.Start( processStartInfo );
process.OutputDataReceived += CaptureOutput;
process.ErrorDataReceived += CaptureOutput;
Thread runThread = new Thread( this.runThread );
runThread.Start( process );
}
else if( (whatToRun & (1 << 3)) != 0 ) // RiskySPN
{
//prg.Show();
prg.progressBar1.Value = 50;
try
{
File.Delete("../../ZBANG/RiskySPN-Master/Results/RiskySPNs-test.csv");
}
catch { }
logString = "\n--------------------------------------------\nPowershell - RiskySPN:\n--------------------------------------------\n\n";
File.AppendAllText( /*"..\\..\\ZBANG\\log.txt"*/logFileName, logString);
prg.label1.Content = "Now running RiskySPN test...";
totalString += "\n\n\n" + DateTime.Now.ToString() + "> Now Running RiskySPN\n";
var processStartInfo = new ProcessStartInfo
{
FileName = @"Powershell.exe",
//Arguments = "-ExecutionPolicy Bypass -noprofile -command \"./Get-PotentiallyCrackableAccounts.ps1; Report-PotentiallyCrackableAccounts -Type 'CSV' -DoNotOpen -Path 'Results/' -Name 'RiskySPNs-test'\"",
//Arguments = "-ExecutionPolicy Bypass -noprofile -Command \"& { .\\Get-PotentiallyCrackableAccounts.ps1; Report-PotentiallyCrackableAccounts -Type CSV -DoNotOpen -Path Results/ -Name RiskySPNs-test}\"",
//Arguments = "-ExecutionPolicy Bypass -noprofile -Command \"& { ./Get-PotentiallyCrackableAccounts.ps1 }",
//##### 26/12/2017 Arguments = "-ExecutionPolicy Bypass -noprofile -command \"Import-Module './RiskySPNs.psm1' -force\"",
Arguments = "-ExecutionPolicy Bypass -noprofile -command \"& { ./Find-PotentiallyCrackableAccounts.ps1 }\"",
//string args = ;
/*-Type 'CSV' -Path '..\\..\\ZBANG\\RiskySPN-master\\Results\\' -DoNotOpen -Name 'RiskySPNs-test'",*/
WorkingDirectory = "../../ZBANG/RiskySPN-Master",
RedirectStandardOutput = true,
UseShellExecute = false,
CreateNoWindow = true
};
var process = Process.Start( processStartInfo );
process.OutputDataReceived += CaptureOutput;
process.ErrorDataReceived += CaptureOutput;
Thread runThread = new Thread( this.runThread );
runThread.Start( process );
}
else if( (whatToRun & (1 << 4)) != 0 )
{
//prg.Show();
prg.label1.Content = "Now running Mystique test...";
totalString += "\n\n\n" + DateTime.Now.ToString() + "> Now Running Mystique\n";
logString = "\n--------------------------------------------\nPowershell - Mistique:\n--------------------------------------------\n\n";
File.AppendAllText( /*"..\\..\\ZBANG\\log.txt"*/logFileName, logString);
try
{
File.Delete("../../ZBANG/Mystique-Master/Results/delegation_info.csv");
}
catch { }
var processStartInfo = new ProcessStartInfo
{
FileName = @"Powershell.exe",
Arguments = "-noprofile -ExecutionPolicy Bypass -command \"& { ./Mystique.ps1 }\"",
WorkingDirectory = "../../ZBANG/Mystique-Master",
RedirectStandardOutput = true,
UseShellExecute = false,
CreateNoWindow = true
};
var process = Process.Start( processStartInfo );
process.OutputDataReceived += CaptureOutput;
process.ErrorDataReceived += CaptureOutput;
Thread runThread = new Thread( this.runThread );
runThread.Start( process );
}
} // endfunc
/**
* @brief Runs powershell within C# to get the different tools running
*
* @param string args - arguments to run
*
**/
void windowRunPowerShell(string args)
{
using (PowerShell PowerShellInstance = PowerShell.Create())
{
Directory.SetCurrentDirectory("../../ZBANG/ACLight-master");
// use "AddScript" to add the contents of a script file to the end of the execution pipeline.
// use "AddCommand" to add individual commands/cmdlets to the end of the execution pipeline.
PowerShellInstance.AddScript( "Set-ExecutionPolicy Bypass");
string script = File.ReadAllText( "ACLight.ps1");
PowerShellInstance.AddScript( script);
//PowerShellInstance.AddCommand("Import-Module").AddArgument("ACLight.psm1");
Collection<PSObject> PSOutput = PowerShellInstance.Invoke();
string outputlog = "-------------------------------------------\n";
foreach (PSObject outputItem in PSOutput)
{
}
PowerShellInstance.Commands.Clear();
PowerShellInstance.AddCommand("Start-ACLsAnalysis");
IAsyncResult result = PowerShellInstance.BeginInvoke();
// do something else until execution has completed.
// this could be sleep/wait, or perhaps some other work
while (result.IsCompleted == false)
{
Console.Write(".");
Thread.Sleep(1000);
// might want to place a timeout here...
}
Console.WriteLine("Finished!");
outputlog += "-------------------------------------------\n";
File.AppendAllText( /*"..\\..\\ZBANG\\log.txt"*/logFileName, "\n\nzBang Launched at " + DateTime.Now.ToString() + "\n" + outputlog + "\n");
} // endusing powershell
} // endfunc run powershell
void CaptureOutput(object sender, DataReceivedEventArgs e)
{
Dispatcher.BeginInvoke(new Action(delegate
{
totalString += e.Data + "\n";
prg.textBlockPowershell.Text += e.Data + "\n";
prg.textBlockPowershell.ScrollToEnd();
File.AppendAllText( /*"..\\..\\ZBANG\\log.txt"*/logFileName, e.Data + "\n");
prg.textBlockPowershell.Select (prg.textBlockPowershell.Text.Length, 0);
prg.myScroll.LineDown();
prg.myScroll.LineDown();
prg.myScroll.LineDown();
prg.myScroll.LineDown();
prg.myScroll.LineDown();
prg.textBlockPowershell.ScrollToEnd();
}));
}
/// <summary>
///
/// </summary>
///
string logString = "";
public void runThread(object process)
{
Process prc = (Process)process;
prc.BeginOutputReadLine();
//prc.BeginErrorReadLine();
while( !prc.HasExited )
{
if( quitApplication )
{
prc.Close();
Dispatcher.BeginInvoke( new Action( delegate
{
//prg.Close();
} ) );
return;
}
}
//prc.WaitForExit();
Dispatcher.BeginInvoke( new Action( delegate
{
//prg.Hide();
string output = "";
/*
try
{
output = prc.StandardOutput.ReadToEnd();
}
catch
{
}
*/
if( (whatToRun & (1 << 0)) != 0 )
{
whatToRun = whatToRun & (~1);
}
else if( (whatToRun & (1 << 1)) != 0 )
{
whatToRun = whatToRun & (~2);
//logString += "Powershell - Skeleton Key:\n" + output + "\n\n\n";
}
else if( (whatToRun & (1 << 2)) != 0 )
{
whatToRun = whatToRun & (~4);
//logString += "Powershell - SID History:\n" + output + "\n\n\n";
}
else if( (whatToRun & (1 << 3)) != 0 )
{
whatToRun = whatToRun & (~(1 << 3));
//logString += "Powershell - RiskySPN:\n" + output + "\n\n\n";
}
else if( (whatToRun & (1 << 4)) != 0 )
{
whatToRun = whatToRun & (~(1 << 4));
//logString += "Powershell - Mystique:\n" + output + "\n\n\n";
}
if( whatToRun == 0 ) whatToRun--;
LaunchZbang.RaiseEvent( new RoutedEventArgs( ButtonBase.ClickEvent ) );
} ) );
return;
#if zero
if( (bool)toggleSkeleton.IsChecked )
{
prg.Show();
prg.label1.Content = "Now running Skeleton Key test...";
var processStartInfo = new ProcessStartInfo
{
FileName = @"Powershell.exe",
Arguments = "-ExecutionPolicy bypass ../../ZBANG/SkeletonKey_Scanner/SkeletonCheck.ps1",
RedirectStandardOutput = true,
UseShellExecute = false
};
var process = Process.Start( processStartInfo );
var output = process.StandardOutput.ReadToEnd();
process.WaitForExit();
prg.Hide();
System.Windows.MessageBox.Show( output, "Powershell - SkeletonKey" );
}
if( (bool)toggleSIDHistory.IsChecked )
{
prg.Show();
prg.label1.Content = "Now running SID History test...";
var processStartInfo = new ProcessStartInfo
{
FileName = @"Powershell.exe",
Arguments = "-noprofile -ExecutionPolicy Bypass Import-Module '../../ZBANG/SIDHistory/SIDHistory_Scanner.ps1' -force ; Report-UsersWithSIDHistory -Type 'CSV' -Path '../../ZBANG/SIDHistory/Results/' -DoNotOpen",
RedirectStandardOutput = true,
UseShellExecute = false
};
var process = Process.Start( processStartInfo );
var output = process.StandardOutput.ReadToEnd();
process.WaitForExit();
prg.Hide();
System.Windows.MessageBox.Show( output, "Powershell - SID History" );
}
if( (bool)toggleRisky.IsChecked )
{
prg.Show();
prg.label1.Content = "Now running RiskySPN test...";
var processStartInfo = new ProcessStartInfo
{
FileName = @"Powershell.exe",
Arguments = "-noprofile -ExecutionPolicy Bypass Import-Module '../../ZBANG/RiskySPN-Master/Get-PotentiallyCrackableAccounts.ps1' -force ; Report-PotentiallyCrackableAccounts -Type 'CSV' -DoNotOpen -Path '../../ZBANG/RiskySPN-master/Results' -Name 'RiskySPNs-test'",
/*-Type 'CSV' -Path '..\\..\\ZBANG\\RiskySPN-master\\Results\\' -DoNotOpen -Name 'RiskySPNs-test'",*/
RedirectStandardOutput = true,
UseShellExecute = false
};
var process = Process.Start( processStartInfo );
var output = process.StandardOutput.ReadToEnd();
process.WaitForExit();
prg.Hide();
System.Windows.MessageBox.Show( output, "Powershell - RiskySPN" );
}
if( (bool)toggleMystique.IsChecked )
{
prg.Show();
prg.label1.Content = "Now running Mystique test...";
var processStartInfo = new ProcessStartInfo
{
FileName = @"Powershell.exe",
Arguments = "-noprofile -ExecutionPolicy Bypass Import-Module '../../ZBANG/Mystique-Master/Mystique.ps1' -force ; Find-DelegationAccounts",
RedirectStandardOutput = true,
UseShellExecute = false
};
var process = Process.Start( processStartInfo );
var output = process.StandardOutput.ReadToEnd();
process.WaitForExit();
prg.Hide();
System.Windows.MessageBox.Show( output, "Powershell - Mystique" );
}
#endif
}
/**
* @brief Called when the application window is quitting
*
**/
private void DataWindow_Closing(object sender, System.ComponentModel.CancelEventArgs e)
{
quitApplication = true;
System.Environment.Exit( 0 );
}
private void Export_clicked(object sender, RoutedEventArgs e)
{
//string getfile = null;
/*
OpenFileDialog openFileDialog = new OpenFileDialog();
if( openFileDialog.ShowDialog() == true )
getfile = openFileDialog.FileName;
else return;
*/
string zipname = "../../../../ZBANG" + (DateTime.Now.Year - 2000).ToString("D02") + DateTime.Now.Month.ToString("D02") + DateTime.Now.Day.ToString("D02") + "-" + DateTime.Now.Hour.ToString("D02") + DateTime.Now.Minute.ToString("D02") + ".zip";
if( File.Exists( zipname ) )
{
MessageBoxResult mbr = MessageBox.Show( "Export file exists in directory.\nThis will overwrite it. Are you sure?", "Warning", MessageBoxButton.YesNo, MessageBoxImage.Question );
if( mbr == MessageBoxResult.No )
return;
File.Delete( zipname );
}
{
//extract the contents of the file we created
//ZipFile.ExtractToDirectory( "../../zbang.zip", "../../ZBANG");
// move the scanner.zip file to another location, zip everything and then replace it back to its location
try
{
File.Copy("../../ZBANG/SkeletonKey_Scanner/scanner.zip", "../../scanner.zip");
File.Delete("../../ZBANG/SkeletonKey_Scanner/scanner.zip");
}
catch {
MessageBox.Show("v0.27 cannot move scanner.zip file...");
}
ZipFile.CreateFromDirectory("../../Zbang", zipname);
try
{
File.Copy("../../scanner.zip", "../../ZBANG/SkeletonKey_Scanner/scanner.zip");
File.Delete("../../scanner.zip");
}
catch {
MessageBox.Show("v0.27 cannot move scanner.zip file BACK...");
}
MessageBox.Show( "Zbang File have been exported successfuly", "Info", MessageBoxButton.OK, MessageBoxImage.Information );
}
}
private void Import_clicked(object sender, RoutedEventArgs e)
{
MessageBoxResult mbr = MessageBox.Show( "This will overwrite current zBang Data. Are you sure?", "Warning", MessageBoxButton.YesNo, MessageBoxImage.Question );
if (mbr == MessageBoxResult.No)
{
runLaunchWindow();
return;
}
string getfile = null;
OpenFileDialog openFileDialog = new OpenFileDialog();
openFileDialog.Title = "Select Import File";
openFileDialog.Filter = "zBang Files|*.zip";
if( openFileDialog.ShowDialog() == true )
getfile = openFileDialog.FileName;
else return;
try
{
using (System.IO.Compression.ZipArchive zip1 = ZipFile.OpenRead(getfile))
{
// here, we extract every entry, but we could extract conditionally
// based on entry name, size, date, checkbox status, etc.
foreach (ZipArchiveEntry file in zip1.Entries)
{
if (file.Name == "")
{// Assuming Empty for Directory
continue;
}
file.ExtractToFile("../../ZBANG/" + file.FullName, true);
}
}
// extract the contents of the file we created
//ZipFile.ExtractToDirectory( getfile, "../../ZBANG" );
}
catch (Exception ex)
{
MessageBox.Show(ex.Message);
}
// NS 03012018 ... . . . . . runLaunchWindow();
toggleACLight.IsChecked = true;
toggleSkeleton.IsChecked = true;
toggleSIDHistory.IsChecked = true;
toggleRisky.IsChecked = true;
toggleMystique.IsChecked = true;
this.Launch.RaiseEvent(new RoutedEventArgs(ButtonBase.ClickEvent));
} // endclass
private void helpButton_clicked(object sender, RoutedEventArgs e)
{
if (rtbACL.Visibility == Visibility.Visible)
{
rtbACL.Visibility = Visibility.Hidden;
return;
}
if (this.ViewModel.onScreen != MainWindowViewModel.ON_SCREEN.ACLLIGHT_ON_SCREEN &&
this.ViewModel.onScreen != MainWindowViewModel.ON_SCREEN.DOMAIN_CARDS &&
this.ViewModel.onScreen != MainWindowViewModel.ON_SCREEN.CARDS_ON_SCREEN_WITH_DOMAIN_SELECTION &&
this.ViewModel.onScreen != MainWindowViewModel.ON_SCREEN.CARDS_ON_SCREEN_NO_DOMAIN_SELECTION)
return;
string curDir = Directory.GetCurrentDirectory();
StreamReader sr = File.OpenText(String.Format("{0}/../../ZBANG/ACLight Attack Path Update.html", curDir));
TextBox.Text = sr.ReadToEnd();
sr.Close();
rtbACL.Visibility = Visibility.Visible;
}
private void Window_Loaded(object sender, RoutedEventArgs e)
{
//zoomControl.ZoomXLoc = this.Width - 100;
//MessageBox.Show("testing123");
/* This has been removed at 01_01 22/11/2018
LicenseWindow eula = new LicenseWindow();
bool rsult = (bool)eula.ShowDialog();
if( !rsult)
{
System.Windows.Application.Current.Shutdown();
return;
}
*/
runLaunchWindow();
}
private void Window_SizeChanged(object sender, SizeChangedEventArgs e)
{
if( e.WidthChanged )
zoomControl.ZoomXLoc = e.NewSize.Width - 100;
if( e.HeightChanged )
zoomControl.ZoomYLoc = e.NewSize.Height - 420;
}
//! WHEN MAGNIFYING GLASS IS CLICKED
private void magnifyingGlassClicked(object sender, MouseButtonEventArgs e)
{
if (popupPicture.IsOpen)
{
popupPicture.IsOpen = false;
return;
}
Image img = (Image)sender;
Person person = (Person)img.DataContext;
//person.Avatar =
popupPicture.IsOpen = true;
popupPicture.PlacementTarget = (UIElement)sender;
imgPopup.Source = person.thumbnail;
// now, somehow show picture....
}
private void relaunch_clicked(object sender, RoutedEventArgs e)
{
runLaunchWindow();
}
} // endclass
//*****************************************************************************************************************
// from codeprpoject: https://www.codeproject.com/Articles/1097390/Displaying-HTML-in-a-WPF-RichTextBox
//*****************************************************************************************************************
public class HtmlRichTextBoxBehavior : DependencyObject
{
public static readonly DependencyProperty TextProperty =
DependencyProperty.RegisterAttached( "Text", typeof( string ),
typeof( HtmlRichTextBoxBehavior ), new UIPropertyMetadata( null, OnValueChanged ) );
public static string GetText(RichTextBox o) { return (string)o.GetValue( TextProperty ); }
public static void SetText(RichTextBox o, string value) { o.SetValue( TextProperty, value ); }
private static void OnValueChanged(DependencyObject dependencyObject,
DependencyPropertyChangedEventArgs e)
{
var richTextBox = (RichTextBox)dependencyObject;
var text = (e.NewValue ?? string.Empty).ToString();
var xaml = HtmlToXamlConverter.ConvertHtmlToXaml( text, true );
var flowDocument = XamlReader.Parse( xaml ) as FlowDocument;
HyperlinksSubscriptions( flowDocument );
richTextBox.Document = flowDocument;
}
private static void HyperlinksSubscriptions(FlowDocument flowDocument)
{
if( flowDocument == null ) return;
GetVisualChildren( flowDocument ).OfType<Hyperlink>().ToList()
.ForEach( i => i.RequestNavigate += HyperlinkNavigate );
}
private static IEnumerable<DependencyObject> GetVisualChildren(DependencyObject root)
{
foreach( var child in LogicalTreeHelper.GetChildren( root ).OfType<DependencyObject>() )
{
yield return child;
foreach( var descendants in GetVisualChildren( child ) ) yield return descendants;
}
}
private static void HyperlinkNavigate(object sender,
System.Windows.Navigation.RequestNavigateEventArgs e)
{
Process.Start( new ProcessStartInfo( e.Uri.AbsoluteUri ) );
e.Handled = true;
}
} // endclass
public static class GetDotNetVersion
{
public static bool Get45PlusFromRegistry( string logFileName)
{
const string subkey = @"SOFTWARE\Microsoft\NET Framework Setup\NDP\v4\Full\";
using( RegistryKey ndpKey = RegistryKey.OpenBaseKey( RegistryHive.LocalMachine, RegistryView.Registry32 ).OpenSubKey( subkey ) )
{
if( ndpKey != null && ndpKey.GetValue( "Release" ) != null )
{
Console.WriteLine( ".NET Framework Version: " + CheckFor45PlusVersion( (int)ndpKey.GetValue( "Release" ) ) );
File.AppendAllText( logFileName, ".NET Framework Version: " + CheckFor45PlusVersion( (int)ndpKey.GetValue( "Release" ) ) );
return true;
}
else
{
Console.WriteLine( ".NET Framework Version 4.5 or later is not detected." );
File.AppendAllText( logFileName, ".NET Framework Version 4.5 or later is not detected." );
return false;
}
}
} // endfunc
// Checking the version using >= will enable forward compatibility.
private static string CheckFor45PlusVersion(int releaseKey)
{
if( releaseKey >= 460798 )
return "4.7 or later";
if( releaseKey >= 394802 )
return "4.6.2";
if( releaseKey >= 394254 )
{
return "4.6.1";
}
if( releaseKey >= 393295 )
{
return "4.6";
}
if( (releaseKey >= 379893) )
{
return "4.5.2";
}
if( (releaseKey >= 378675) )
{
return "4.5.1";
}
if( (releaseKey >= 378389) )
{
return "4.5";
}
// This code should never execute. A non-null release key should mean
// that 4.5 or later is installed.
return "No 4.5 or later version detected";
}
}
// This example displays output like the following:
// .NET Framework Version: 4.6.1
} // endnamespace
|
In a lab just outside Boston, Tufts University professor and engineer Mike Zimmerman and his team have created what might be the next generation lithium-ion battery that will safely power our phones, cars, and more.
The key to the invention is a solid plastic electrolyte, the substance that bridges the gap between the positive and negative electrodes. In most lithium-ion batteries, the electrolyte is a liquid, and that can make them vulnerable to fire or explosion when hit or pierced. Those vulnerabilities were recently on display in the recalled Samsung Galaxy Note 7 phones, where the battery would spontaneously explode or catch fire because a corner of the battery casing was too small, bending one of the electrodes and increasing the risk of short circuits.
But Zimmerman’s battery can withstand repeated damage without risking explosion or fire. In fact, it can continue to power devices even after most of it has been chopped away. |
White, undecorated paper, not to exceed the size of 8 1/2" X 11", or unstamped white envelopes, including carbon paper and white envelopes with the offender's commitment name and TDCJ number preprinted in the return address portion of the envelope, but excluding any paper with names, addresses or letterhead, and excluding tablets or writing pads with stapled binding. (NOTE: Ruled white paper is not considered decorated and is permitted. Ruled lines can be any color.) Sketch paper is also permitted.
Effective October 1, 2007, offenders are not allowed to receive colored paper from an approved vendor.
Offenders will still be allowed to receive note cards with matching envelopes and journals with white paper. However, offenders will not be allowed to receive the yellow legal pads.
Notecards with matching/colored envelopes are allowed.
Rest assured that the stationary products offered on our site meet these standards. |
india
Updated: Jun 19, 2019 12:17 IST
:A seven-year-old girl was killed and at least nine people received injuries in three incidents of roof collapse in Fatehabad and Hisar districts on Tuesday in the wake of intermittent rain in the last three days.
Eight members of a family from Bhattu village in Fatehabad district were buried under the debris when the roof of their house collapsed around 6am. The injured were rushed to the Fatehabad civil hospital but Mohini (7) succumbed to her injuries during treatment.
Bhattu village Sarpanch Roshan Singh said the house belonging to Darshan Singh was constructed nearly 20 years ago. Darshan said all members of the family were sleeping in the same room. Other injured were identified as Nitu (32), Manju (20), Mesar Devi (60), Aarti (7), Ayan (10), Pooja (10) and Ajay (10).
Police sent the girl’s body for post-mortem to the Fatehabad general hospital and started investigation under Section 174 of the Code of Criminal Procedure (CrPC).
In Bhuna town of the district, a 34-year-old man, Mahipal Singh, was injured after the roof of his house collapsed about 6.30am.He was rushed to the local community health centre (CHC), Bhuna. Mahipal said he was getting ready to go to fields when the roof of his verandah collapsed. Police reached the spot and enquired about the incident.
At Kinala village of Hisar district, a 40-year-old woman, Darshana Devi, received minor injuries after the roof of her house collapsed about 9am. Locals rushed to the spot to rescue her. Darshana said there was no one else in the house when the incident took place. She was taken to a private hospital for treatment.
Police spokesperson Harish Bhardwaj said, “We got information about the incident and a police team from Uklana block reached the spot.”
The meteorological department at the Chaudhary Charan Singh Haryana Agricultural University (CCSHAU) said Hisar district recorded 64.4mm rainfall in the last three days while Fatehabad district received 14.3mm rain in the last 24 hours. |
Q:
R map switzerland according to NPA (locality)
I plan to do a survey in Switzerland. NPA will be asked.
NPA (postal codes) contains 4 number.
For instance 1227 is the NPA of Carouge (part of canton Geneva - Switzerland).
For instance 1784 is the NPA of Courtepin (part of canton Fribourg -Switzerland).
etc.
I would like to know how to represent all observation (about 1500) on a map. I was thinking using ggplot as I use it for other graphs (I think ggplot is "beautiful"). However, I'm open to any other suggestion.
Here are some fake data:
http://pastebin.com/HsuQnLP3
The output for the swiss map should be a bit like that USA map (credit:http://www.openintro.org)
Update:
I've tried to create some code :
library(sp)
test <- url("https://dl.dropboxusercontent.com/u/6421260/CHE_adm3.RData")
print(load(test))
close(test)
gadm$NAME_3
gadm$TYPE_3
But it seems http://gadm.org/ doesn't provide the NPA of the communes...
New update:
I've find (thanks @yrochat) a shapefile with NPA:
http://www.cadastre.ch/internet/cadastre/fr/home/products/plz/data.html
I'ts the ZIP file called : Shape LV03
Then I've tried
library("maptools")
swissmap <- readShapeLines("C:/Users/yourName/YourPath/PLZO_SHP_LV03/PLZO_PLZ.shp")
plot(swissmap)
data <- data.frame(swissmap)
data$PLZ #the row who gives the NPA
As I have the PLZ on a shapefile, how do i color my observation on the map?
I provided some fake data on data http://pastebin.com/HsuQnLP3
Thanks
A:
OK, with the shapefile, we can plot things easily enough.
work.dir <- "directory_name_no_trailing slash"
# open the shapefile
require(rgdal)
require(rgeos)
require(ggplot2)
ch <- readOGR(work.dir, layer = "PLZO_PLZ")
# convert to data frame for plotting with ggplot - takes a while
ch.df <- fortify(ch)
# generate fake data and add to data frame
ch.df$count <- round(runif(nrow(ch.df), 0, 100), 0)
# plot with ggplot
ggplot(ch.df, aes(x = long, y = lat, group = group, fill = count)) +
geom_polygon(colour = "black", size = 0.3, aes(group = group)) +
theme()
# or you could use base R plot
ch@data$count <- round(runif(nrow(ch@data), 0, 100), 0)
plot(ch, col = ch@data$count)
Personally I find ggplot a lot easier to work with than plot and the default output is much better looking.
And ggplot uses a straightforward data frame, which makes it easy to subset.
# plot just a subset of NPAs using ggplot
my.sub <- ch.df[ch.df$id %in% c(4,6), ]
ggplot(my.sub, aes(x = long, y = lat, group = group, fill = count)) +
geom_polygon(colour = "black", size = 0.3, aes(group = group)) +
theme()
Result:
|
In a mobile communication system such as IMT-2000, transmission power control is performed from the viewpoint of enlargement of circuit capacity and economy of battery of a mobile station and the like. For example, quality measurement of a channel is performed in a receiving side, and a transmission power control (TPC) bit is transmitted by a return channel (DPCCH, for example) such that the channel that is being received satisfies desired quality. As a result, the transmission power is updated by 1 dB, for example, and quality measurement and transmitting/receiving of the TPC bit are repeated, so that the transmission power can be gradually changed to be closer to an optimal value. That is, in a communication of a circuit switching scheme, an individual channel is assigned specifically to a mobile station, and the transmission power of the mobile station is gradually adjusted based on a temporally continuing history on the transmission power. Such transmission power control is described in a non-patent document 1, for example. [Non Patent document 1] Keiji Tachikawa, “W-CDMA mobile communication scheme”, MARUZEN, pp. 126-128 |
// Copyright 2015 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package precis
import (
"bytes"
"fmt"
"reflect"
"testing"
"golang.org/x/text/internal/testtext"
"golang.org/x/text/transform"
)
type testCase struct {
input string
output string
err error
}
func doTests(t *testing.T, fn func(t *testing.T, p *Profile, tc testCase)) {
for _, g := range enforceTestCases {
for i, tc := range g.cases {
name := fmt.Sprintf("%s:%d:%+q", g.name, i, tc.input)
testtext.Run(t, name, func(t *testing.T) {
fn(t, g.p, tc)
})
}
}
}
func TestString(t *testing.T) {
doTests(t, func(t *testing.T, p *Profile, tc testCase) {
if e, err := p.String(tc.input); tc.err != err || e != tc.output {
t.Errorf("got %+q (err: %v); want %+q (err: %v)", e, err, tc.output, tc.err)
}
})
}
func TestBytes(t *testing.T) {
doTests(t, func(t *testing.T, p *Profile, tc testCase) {
if e, err := p.Bytes([]byte(tc.input)); tc.err != err || string(e) != tc.output {
t.Errorf("got %+q (err: %v); want %+q (err: %v)", string(e), err, tc.output, tc.err)
}
})
t.Run("Copy", func(t *testing.T) {
// Test that calling Bytes with something that doesn't transform returns a
// copy.
orig := []byte("hello")
b, _ := NewFreeform().Bytes(orig)
if reflect.ValueOf(b).Pointer() == reflect.ValueOf(orig).Pointer() {
t.Error("original and result are the same slice; should be a copy")
}
})
}
func TestAppend(t *testing.T) {
doTests(t, func(t *testing.T, p *Profile, tc testCase) {
if e, err := p.Append(nil, []byte(tc.input)); tc.err != err || string(e) != tc.output {
t.Errorf("got %+q (err: %v); want %+q (err: %v)", string(e), err, tc.output, tc.err)
}
})
}
func TestStringMallocs(t *testing.T) {
if n := testtext.AllocsPerRun(100, func() { UsernameCaseMapped.String("helloworld") }); n > 0 {
// TODO: reduce this to 0.
t.Skipf("got %f allocs, want 0", n)
}
}
func TestAppendMallocs(t *testing.T) {
str := []byte("helloworld")
out := make([]byte, 0, len(str))
if n := testtext.AllocsPerRun(100, func() { UsernameCaseMapped.Append(out, str) }); n > 0 {
t.Errorf("got %f allocs, want 0", n)
}
}
func TestTransformMallocs(t *testing.T) {
str := []byte("helloworld")
out := make([]byte, 0, len(str))
tr := UsernameCaseMapped.NewTransformer()
if n := testtext.AllocsPerRun(100, func() {
tr.Reset()
tr.Transform(out, str, true)
}); n > 0 {
t.Errorf("got %f allocs, want 0", n)
}
}
func min(a, b int) int {
if a < b {
return a
}
return b
}
// TestTransformerShortBuffers tests that the precis.Transformer implements the
// spirit, not just the letter (the method signatures), of the
// transform.Transformer interface.
//
// In particular, it tests that, if one or both of the dst or src buffers are
// short, so that multiple Transform calls are required to complete the overall
// transformation, the end result is identical to one Transform call with
// sufficiently long buffers.
func TestTransformerShortBuffers(t *testing.T) {
srcUnit := []byte("a\u0300cce\u0301nts") // NFD normalization form.
wantUnit := []byte("àccénts") // NFC normalization form.
src := bytes.Repeat(srcUnit, 16)
want := bytes.Repeat(wantUnit, 16)
const long = 4096
dst := make([]byte, long)
// 5, 7, 9, 11, 13, 16 and 17 are all pair-wise co-prime, which means that
// slicing the dst and src buffers into 5, 7, 13 and 17 byte chunks will
// fall at different places inside the repeated srcUnit's and wantUnit's.
if len(srcUnit) != 11 || len(wantUnit) != 9 || len(src) > long || len(want) > long {
t.Fatal("inconsistent lengths")
}
tr := NewFreeform().NewTransformer()
for _, deltaD := range []int{5, 7, 13, 17, long} {
loop:
for _, deltaS := range []int{5, 7, 13, 17, long} {
tr.Reset()
d0 := 0
s0 := 0
for {
d1 := min(len(dst), d0+deltaD)
s1 := min(len(src), s0+deltaS)
nDst, nSrc, err := tr.Transform(dst[d0:d1:d1], src[s0:s1:s1], s1 == len(src))
d0 += nDst
s0 += nSrc
if err == nil {
break
}
if err == transform.ErrShortDst || err == transform.ErrShortSrc {
continue
}
t.Errorf("deltaD=%d, deltaS=%d: %v", deltaD, deltaS, err)
continue loop
}
if s0 != len(src) {
t.Errorf("deltaD=%d, deltaS=%d: s0: got %d, want %d", deltaD, deltaS, s0, len(src))
continue
}
if d0 != len(want) {
t.Errorf("deltaD=%d, deltaS=%d: d0: got %d, want %d", deltaD, deltaS, d0, len(want))
continue
}
got := dst[:d0]
if !bytes.Equal(got, want) {
t.Errorf("deltaD=%d, deltaS=%d:\ngot %q\nwant %q", deltaD, deltaS, got, want)
continue
}
}
}
}
|
President Barack Obama speaks at Arlington National Cemetery in Virginia on May 30, 2016.
Photographer: Mike Theiler-Pool/Getty Images
President Barack Obama will return Wednesday to the site of his first presidential trip, an Indiana city buoyed by a resurgent motor home industry, for an economic victory lap and to argue for a Democratic successor.
Elkhart, Indiana, has seen its unemployment rate drop from 18.9 percent in 2009, Obama’s first year in office, to just 4.1 percent in April. The White House views the economic rebound of the 51,000-population city as a frame for the success of programs such as Obama’s stimulus and his health-care law.
Obama plans to use the visit to argue that only a Democratic successor can build on policies that saved cities like Elkhart from economic demise. Both the presumptive Republican presidential nominee, Donald Trump, and his likely opponent, Democrat Hillary Clinton, have identified Rust Belt voters as a key constituency in the 2016 presidential election.
“If what you really care about in this election is your pocketbook; if what you’re concerned about is who will look out for the interests of working people and grow the middle class, then the debate isn’t even close,” Obama will say in a speech at Elkhart, according to excerpts released by the White House.
‘Working Families’
Keep up with the best of Bloomberg Politics.
Get our newsletter daily.
Business
Your guide to the most important business stories of the day, every day.
You will now receive the Business newsletter
Markets
The most important market news of the day. So you can sleep an extra five minutes.
You will now receive the Markets newsletter
Technology
Insights into what you'll be paying for, downloading and plugging in tomorrow and 10 years from now.
You will now receive the Technology newsletter
Pursuits
What to eat, drink, wear and drive – in real life and your dreams.
You will now receive the Pursuits newsletter
Game Plan
The school, work and life hacks you need to get ahead.
You will now receive the Game Plan newsletter
“One path would lower wages, eliminate worker protections, cut investments in things like education, weaken the safety net, kick people off health insurance, and let China write the rules for the global economy,” Obama will say. “It would let big oil weaken rules that protect our air and water, and big banks weaken rules that protect families from getting cheated, and cut taxes for the wealthiest Americans to historic lows.”
Democrats’ policies, by contrast, would raise wages, improve employee benefits and result in a “fairer tax code,” the president will argue.
“It will make a real difference for the prospects of working families,” he will say. “It will grow the middle class.”
Elkhart suffered during the recession when employees were fired at the city’s RV manufacturers -- companies there include Thor Industries Inc., Nexus RV and Berkshire Hathaway Inc. subsidiary Forest River -- and after the 2006 closing of a Bayer AG plant that made Alka-Seltzer and Flintstones vitamins.
But a recovery in the RV industry, which has benefited from low gas prices, has led to an economic rebound.
Hiring Surge
Manufacturing jobs in the city, a five-minute drive from the Michigan border, have more than doubled from 12,000 in 2009 to 26,000 today, though they remain below historical highs. Federally funded projects to repair the city’s streets and repave the airport’s runway put construction workers back on the job.
Graduation rates at the county’s public high schools have jumped from 75 percent to nearly 90 percent, and the number of homes in the process of foreclosure has dropped from 9.5 percent in 2010 to 3.7 percent today.
Obama isn’t expected to explicitly discuss the presidential candidates, or break his silence on the Democratic nomination fight between Clinton and Vermont Senator Bernie Sanders. But in the excerpts of his remarks, he implicitly criticized Trump, who has promised to build a wall on the border with Mexico and prohibit immigration by Muslims.
Blue-Collar Voters
“One thing I can promise you is that if we turn against each other based on divisions of race or religion, then we won’t build on the progress we’ve started,” Obama will say. “If we get cynical and just vote our fears, or don’t vote at all, we won’t build on the progress we’ve started.”
Indiana is home to many of the white, blue-collar voters gravitating to the populist campaign of Donald Trump, who are frustrated that Obama’s economic recovery overlooked them and are concerned that globalization threatens their livelihoods. Democratic and Republican voters both said the economy was their top issue according to exit polls during the state’s primary election last month.
That fear is particularly acute among some of the state’s biggest businesses. Statewide, nearly 30 percent of the gross domestic product comes from manufacturing, primarily cars and steel, according to the U.S. Bureau of Economic Analysis.
Greater Tariffs
Trump has repeatedly seized on the erosion of U.S. manufacturing in his campaign, threatening to impose greater tariffs on foreign trading partners. He has said he would impose additional taxes on Carrier-brand air conditioners manufactured in Mexico, after the United Technologies Corp. subsidiary shuttered an Indiana plant and moved its operations abroad.
“I wanna do the number on Carrier, folks,” Trump said in April. “I don’t like what they did.”
Senate Majority Leader Mitch McConnell, a Kentucky Republican, questioned Obama’s economic message, saying in an interview with CNBC Wednesday that the president has harmed economic growth because he’s “clogged up the whole system” with regulations.
“The average American is worse off than they were at the end of ’07 and considerably worse off than they were when the president took office,” McConnell said.
Tax Cuts
And Indiana Governor Mike Pence, a Republican, argued in an op-ed published Wednesday in the city’s local paper, The Elkhart Truth, that the recovery there had happened in spite of the president’s policies, not because of them. Pence credited state-level tax cuts and deregulation for Indiana’s economic improvement.
“We have worked every day to lower the burden of taxes and regulations so businesses large and small can focus on jobs and growth instead of worrying about the burden of their state government,” he wrote.
While middle class incomes have been slow to recover since the recession, they are now slightly higher, even adjusted for inflation. Median household income in the U.S. was $57,243 in April, about 0.6 percent higher than when the recession began in December 2007, according to Annapolis, Maryland-based Sentier Research.
The economy was losing jobs at a rate of 791,000 a month when Obama took office, and unemployment was 7.8 percent. The unemployment rate in April was 5 percent, and the economy added 160,000 jobs.
Industrial Turnaround
The job for Obama is convincing blue-collar voters who have experienced job losses and contraction in their industries that his programs have modernized the economies of states such as Indiana. The White House argues that Elkhart embodies that turnaround.
It’s crucial to Clinton that voters buy Obama’s message in states like Ohio, which the president won by just two percentage points four years ago, and Pennsylvania, where he prevailed by five percentage points. If Trump is able to make inroads in those states, Democrats’ presumed advantage in the Electoral College may evaporate.
Elkhart also holds particular sentimental value to the president, who has repeatedly visited the northern Indiana town throughout his political career.
He campaigned there as a candidate and then as the presumptive Democratic nominee during his first White House bid in 2008. The city was Obama’s first visit as president after Indiana, traditionally a conservative stronghold, narrowly voted for him. Later in 2009, Obama again visited the state and noted that Navistar International Corp., with a factory near Elkhart, had secured a federal grant under the stimulus to develop battery-powered electric vehicles. |
Pac-12 Basketball Recruiting: Arizona and Washington Are in "Good Shape"
What do the Arizona Wildcats and the Washington Huskies men’s basketball programs have in common outside of both being in the Pac-12 conference?
Well, they were both recently listed on the ZagsBlog “12 Schools in Good Shape Heading Into Late July.” The list is composed of programs that Adam Zagoria thinks are ahead of the crowd when it comes to 2013 recruiting at this point in time.
Not only are the Wildcats and Huskies both on that list, but they also share two of the same potential recruits (Aaron Gordon and Jabari Bird), as Zagoria’s reason for their inclusion.
On the recruiting trail, Washington coach Lorenzo Romar was at the Super 64 AAU tournament watching the Dream Vision vs. New Heights NYC game last night. It can be assumed Romar was there along with several other coaches (including UCLA’s Ben Howland) to check out 2013 recruit Isaac Hamilton. Hamilton is currently Rivals’ No. 12 overall recruit in the 2013 class.
Hamilton’s Dream Vision team lost the game, 67-40, and adding insult to injury, his first shot attempt was a three-pointer air ball (h/t Las Vegas Sun). Following the loss, AAU coach Clayton Williams reportedly told his Dream Vision team, “You’re on center stage!...You’re not as good as you think.”
While the AAU team will have other chances to make up for the loss today, the clock is winding down for Hamilton, who recently tweeted, “Last tournament of the year, I’m going to miss AAU”.
Recruiting doesn’t end with Hamilton for Washington, as the program is after several elite recruits in the 2013 class. One of those recruits, Gordon, that both Arizona and Washington have been pursuing, recently had this to say about Romar (via CBSSports' Jeff Borzello):
Stephen Dunn/Getty Images
Romar is one of those coaches that talks with me. He's a floor general. When he was a player, he was a point guard, so he saw two or three plays ahead. I love that in a coach. He has a high basketball IQ. He will adjust his system to his players.
With regard to Arizona coach Sean Miller, Gordon had this to say (h/t Seattle Times’ Percy Allen, via Scout.com’s Michael Luke), “I like Arizona for many of the same reasons I like Washington. Coach Miller is a very good coach and also plays an up-tempo style. Arizona has better facilities, though.”
Both of these programs are competing with last season’s NCAA tournament champion, Kentucky, among others for this top recruit, who likely won’t make a decision until the spring. With Rupp Arena and everything else, the Kentucky Wildcats aren’t hurting for facilities either.
While it can be a little bit of an arms race, Washington’s facilities could use an upgrade. The Seattle Times’ Allen recently quoted Washington’s athletic director as having this to say on the matter:
We're always reviewing our options and trying to make the best possible use of how we source and what we do with Alaska Airlines Arena. It's the whole arena and not just Hec Ed...We're looking at sourcing the whole arena and making it as efficient as possible. We're in the process of looking at that and looking at what are our options going forward, but it's premature to say we're doing anything. What I would say is we're early planning and looking at our options.
It’s becoming an exciting summer in recruiting for both Pac-12 teams, Arizona and Washington. It’s tough to say where these programs will end up once all the dust settles on the 2013 recruit class, but both appear to be in a pretty good position as things currently stand. |
Marine scientists from the University of Queensland have discovered hybridized sharks off Australia´s east coast, leading them to believe that some of these predatory beasts display a tendency to interbreed, challenging long-standing scientific theories regarding shark behavior.
This is the first time scientists have confirmed a substantial number of hybrid sharks off Australia´s coast, speculating that it may be an adaptation or reaction to climate change; and scientists now believe it may indicate that other shark and ray species may interbreed in reaction to climate change.
“Hybridization could enable the sharks to adapt to environmental change as the smaller Australian black tip currently favors tropical waters in the north while the larger common black tip is more abundant in sub-tropical and temperate waters along the south-eastern Australian coastline,” said researcher Jennifer Ovenden of the University of Queensland in a press release.
Ovenden and her colleagues said they had discovered 57 sharks that are a cross between the Australian blacktip shark and the common blacktip shark. Both are genetically distinct species, but are closely related, making it easier for the hybridization to occur.
“Wild hybrids are usually hard to find, so detecting hybrids and their offspring is extraordinary,” Ovenden said. “To find 57 hybrids along 2000km [1240 miles] of coastline is unprecedented.”
The hybridization was confirmed using DNA measurements. The Fisheries Research and Development Corporation co-funded the research, which identified a mismatch between species identification using mitochondrial DNA sequencing and species identification using morphological characters — mature length, birth length, and number of vertebrae.
A nuclear DNA marker was sequenced to confirm the hybridization. Dr. Colin Simpfendorfer of James Cook University´s Fishing and Fisheries Research Center said blacktip sharks were one of the most studied species in tropical Australia. “The results of this research show that we still have a lot to learn about these important ocean predators,” he said.
Scientists from The University of Queensland, James Cook University´s Fishing and Fisheries Research Center, the Queensland Department of Employment, Economic Development and Innovation and the New South Wales Department of Primary Industries are further investigating the full extent of the hybrid zone and are attempting to measure hybrid fitness.
Jess Morgan, another University of Queensland researcher, said that hybrid species are common in fish because their eggs are fertilized in the water. “Sharks physically mate, which is usually a good way to make sure you don’t hybridize with the wrong species,” he told The Australian newspaper.
—
On the Net:
Comments
comments |
The present invention generally relates to matrix materials for use in a wide variety of end use fields and applications. More particularly, the invention relates to new and improved self-repairing, settable or curable matrix material systems containing so-called smart-release fiber reinforcements, alone or in combination with other reinforcement. My prior parent application, Ser. No. 540,191, filed Jun. 19, 1990, describes the new and improved inorganic and organic matrix composites employing concrete matrix systems and asphalt matrix systems as illustrative embodiments. That prior application describes smart-release hollow fiber a additives in settable construction materials and thermoplastic matrices, such as asphalt. This application is being filed to describe other embodiments of the smart-release matrix composite materials generally described in my earlier application and to provide additional examples of end use applications to which these new and improved compositions, articles and methods may be specially adapted and used.
Cement is a fine, gray powder consisting of alumina, lime, silica and iron oxide which sets to a hard material after mixture with water. Cement, along with sand and stone aggregate, make up concrete, the most widely used building material in the world. Steel reinforcing bars (rebars) are commonly added to the interior of concrete for additional strength.
There are many reasons for the popularity of concrete. It is relatively inexpensive, capable of taking on the shape of a mold, has exceptionally high compression strength and is very durable when not exposed to repeated freeze-thaw cycles. However, as a building or construction material, concrete, whether it is reinforced or not, is not without some shortcomings. One major drawback of concrete is that it is relatively low in tensile strength. In other words, it has little ability to bend. Concrete also has little impact resistance and is frequently brittle. A third major drawback is that its durability is significantly reduced when it is used in applications which require it to be exposed to repeated freeze-thaw cycles in the presence of water. Concrete is relatively porous and water is able to permeate the material. Freezing and thawing with the accompanying expansion and contraction of the water, forms cracks in the concrete. Furthermore, if salt is also present in the environment, it dissolves in the water and permeates into the concrete where it is capable of inducing corrosion in any of the rebars or other metallic reinforcements present.
Various techniques have been suggested in the past for overcoming these drawbacks. The addition of fibers to concrete has improved its tensile strength but has decreased its compression strength. Providing exterior coatings on the outer surfaces of the concrete has reduced water permeation, but it is a time-consuming additional step and has little, if any, effect on the lasting strength of the concrete. The addition of modifying agents as freely-mixed additives into a concrete mixture before setting has also been tried. These efforts have met with generally unsatisfactory results. Attempts to add modifying agents in the form of micronodules or prills have also been tried. Frequently, the prills are designed to be heat melted to cause release of the modifying agent into the matrix after setting of the materials. These designs require the application of heat to release the beneficial additive into the matrix after cure. Moreover, the melted, permeated agents leave behind voids in the concrete which weakens the overall structure under load. Accordingly, a demand still exists for an improved concrete matrix material having greater tensile strength, greater durability and comparable or improved compression strength.
In addition to cementitious building materials, the use of polymer composites as structural materials has grown tremendously in recent years. Polymer composite materials have advantages over steel or concrete including good durability, vibration damping, energy absorption, electromagnetic transparency, toughness, control of stiffness, high stiffness to weight ratios, lower overall weight and lower transportation cost. These polymer matrix materials comprise a continuous polymer phase with a fiber reinforcement therein. Some polymer composite materials are three times stronger than steel and five times lighter. They have heretofore been generally more expensive but their use may, in the long term, be economical because of their greatly reduced life cycle costs. Europeans have made bridges completely of specialty polymer matrix composite materials. The polymer composite materials may be used as rebars, tensioning cables, in bonded sheets, wraps, decks, supports, beams or as the primary structures for bridges, decks or buildings. Structures made from polymer matrix materials are specially effective in aggressive environments or are well adapted for building structures where electromagnetic transparency may be needed for highways, radar installations and hospitals.
As used herein, matrix composite materials may refer to generally any continuous matrix phase whether it comprises a settable construction material such as cementitious materials or a thermoplastic material such as asphalt materials, as well as, other synthetic or natural high polymer materials ceramics, metals and other alloy materials. The matrix composite materials include various fiber reinforcements therein distributed throughout the matrix or placed at desired locations within the continuous phase. The matrix composite materials may be fabricated as large building structures and load bearing shaped articles, or they may be molded or machined as small parts for specialty uses. For example, the matrix material may comprise a thin sheet or web of material in the form of a foil, wrap, tape, patch or in strip form. As presently used in this specification, the term matrix composite material does not necessarily refer to large civil engineering structures such as highways and bridges.
In connection with the polymer and/or metal or ceramic matrix composite materials, as well as, in the settable building materials such as concrete materials, special problems cause structures made from these materials to become aged or damaged in use. More particularly, special structural defects arise in use including microcracking, fiber debonding, matrix delamination, fiber breakage, and fiber corrosion, to name but a few. Any one of these microscopic and macroscopic phenomena may lead to failures which alter the strength, stiffness, dimensional stability and life span of the materials. Microcracks, for example, may lead to major structural damage and environmental degradation. The microcracks may grow into larger cracks with time and cause overall material fatigue so that the material deteriorates in long-term use.
Advanced matrix composites used in structural applications are susceptible to damage on both the macro- and microscopic levels. Typical macroscopic damage to composite laminates involves delaminations and destruction of the material due to impact. On the micrographic scale, damage usually involves matrix microcracking and/or debonding at the fiber/matrix interface. Internal damage such as matrix microcracking alters the mechanical properties of shaped articles made therefrom such as strength, stiffness and dimensional stability depending on the material type and the laminate structure. Thermal, electrical and acoustical properties such as conductance, resistance and attenuation have also been shown to change as matrix cracks initiate. Microcracks act as sites for environmental degradation as well as for nucleation of microcracks. Thus, microcracks can ultimately lead to overall material degradation and reduced performance.
Moreover, prior studies have shown that microcracks cause both fiber and matrix dominated properties of the overall composite to be effected. Fiber dominated properties such as tensile strength and fatigue life may be reduced due to redistribution of loads caused by matrix damages. Matrix dominated properties on the other hand such as compressive residual strength may also be influenced by the amount of matrix damage. The impact responses of toughened polymer matrix composites have been studied and it has been shown that matrix cracking precedes delamination which, in turn, precedes fiber fracture. Tough matrices which can reduce or prevent matrix cracking tend to delay the onset of delamination which results in an improved strength composite and longer lasting composite material.
Repair of damages is a major problem when these matrix composite materials are employed in large-scale construction or advanced structures. Macroscale damage due to delamination, microcracking or impacts may be visually detected and can be repaired in the field by hand. Microscale damage occurring within the matrix is likely to go undetected and the damage which results from this type of breakdown may be difficult to detect and very difficult to repair.
In order to overcome the shortcomings of the prior art construction and polymer, ceramic or metal matrix composite materials, it is an object of the present invention to provide new and improved smart structural composite materials having a self-healing capability whenever and wherever cracks are generated.
It is another object of the present invention to provide new and improved composite materials including self-repairing reinforcing fibers capable of releasing chemical agents into the local microscopic domains of the matrix to repair matrix microcracks and rebond damaged interfaces between fibers and matrices.
It is a further object of the present invention to provide a new and improved structural material.
It is another object of the present invention to provide a new and improved cementitious material.
It is still a further object of the present invention to provide a new and improved cementitious or other construction composite material having greater durability and greater tensile strength.
It is still another object of the present invention to provide a new and improved matrix composite materials containing smart self-repairing fiber reinforcement containing repair chemicals therein which may be released by the smart fibers as needed in response to an external stimulus, and optionally which may be refilled with additional repair chemicals as needed in the field.
In accordance with these and other objects, the present invention provides new and improved shaped articles comprising:
a cured matrix material having a plurality of hollow fibers dispersed therein, the hollow fibers having a selectively releasable modifying agent contained therein, means for maintaining the modifying agent within the fibers until selectively released and means for permitting selective release of the modifying agent from the hollow fibers into the matrix material in response to at least one predetermined external stimulus. In accordance with this invention the shaped articles are matrix composite materials of varying size and end use applications. The cured matrix materials have within them smart fibers capable of delivering repair chemicals into the matrix wherever and whenever they are needed.
The present invention also provides a new and improved method for providing shaped articles having long-term durability and environmental degradation resistance comprising the steps of providing a curable matrix composition, distributing a plurality of hollow fibers therein in desired manner so that the hollow fibers are disposed within the matrix material in a desired predetermined distribution. The hollow fibers are filled with a selectively releasable modifying agent therein which is not released during the mixing or distributing step. The fibers are structured so that the modifying agents stay within the interior spaces or cavities of the fibers within the matrix until the matrix is cured or set. After curing, the modifying agents are selectively released from the fibers by application or action of at least one predetermined external stimulus.
In a preferred embodiment, the method of providing a improved durability shaped article comprises providing a cured matrix material containing smart self repair fibers reinforcement therein.
The principles of the present invention are applicable to space age polymer, metal and/or ceramic structural matrix composite materials as well as more conventional cementitious settable or curable building or construction materials.
Other objects and advantages will become apparent from the following Detailed Description of the Preferred Embodiments, taken in conjunction with the Drawings in which: |
Genotypic characterisation of endemic VanA Enterococcus faecium strains isolated in a paediatric hospital.
A total of 36 vancomycin-resistant Enterococcus faecium isolates obtained from 30 patients during a 28-month period in a paediatric university hospital was analysed by pulsed-field gel electrophoresis (PFGE) combined with Southern hybridisation of a vanA-specific DNA probe. All the isolates hybridised with the vanA probe. Seventeen different PFGE patterns and 11 PFGE subtypes were identified among the 36 clinical isolates, and the size of probe-positive bands ranged from c. 30 to 300 kb. These data are consistent with an increase in the overall genomic diversity of vancomycin-resistant E. faecium isolates during the study period. Two periods were distinguished. The prevalence of a single clone in the initial period suggested transmission between patients in three wards. During the following period, multiple genotypes of vancomycin-resistant E. faecium were identified, indicative of multiple introductions or the dissemination of resistance genes by recombinant transposition. |
/**
* iperf-liked network performance tool
*
*/
#include <rtthread.h>
#ifdef PKG_NETUTILS_IPERF
#include <rtdevice.h>
#include <string.h>
#include <stdint.h>
#include <stdio.h>
#include <fcntl.h>
#include <unistd.h>
#include <sys/time.h>
#include <sys/socket.h>
#include <sys/select.h>
#include "netdb.h"
#define IPERF_PORT 5001
#define IPERF_BUFSZ (4 * 1024)
#define IPERF_MODE_STOP 0
#define IPERF_MODE_SERVER 1
#define IPERF_MODE_CLIENT 2
typedef struct
{
int mode;
char *host;
int port;
} IPERF_PARAM;
static IPERF_PARAM param = {IPERF_MODE_STOP, NULL, IPERF_PORT};
static void iperf_udp_client(void *thread_param)
{
int sock;
rt_uint32_t *buffer;
struct sockaddr_in server;
rt_uint32_t packet_count = 0;
rt_uint32_t tick;
int send_size;
send_size = IPERF_BUFSZ > 1470 ? 1470 : IPERF_BUFSZ;
buffer = malloc(IPERF_BUFSZ);
if (buffer == NULL)
{
return;
}
memset(buffer, 0x00, IPERF_BUFSZ);
sock = socket(PF_INET, SOCK_DGRAM, 0);
if(sock < 0)
{
rt_kprintf("can't create socket!! exit\n");
return;
}
server.sin_family = PF_INET;
server.sin_port = htons(param.port);
server.sin_addr.s_addr = inet_addr(param.host);
rt_kprintf("iperf udp mode run...\n");
while (param.mode != IPERF_MODE_STOP)
{
packet_count++;
tick = rt_tick_get();
buffer[0] = htonl(packet_count);
buffer[1] = htonl(tick / RT_TICK_PER_SECOND);
buffer[2] = htonl((tick % RT_TICK_PER_SECOND) * 1000);
sendto(sock, buffer, send_size, 0, (struct sockaddr *)&server, sizeof(struct sockaddr_in));
}
closesocket(sock);
free(buffer);
}
static void iperf_udp_server(void *thread_param)
{
int sock;
rt_uint32_t *buffer;
struct sockaddr_in server;
struct sockaddr_in sender;
int sender_len, r_size;
rt_uint64_t sentlen;
rt_uint32_t pcount = 0, last_pcount = 0;
rt_uint32_t lost, total;
rt_tick_t tick1, tick2;
float f;
char speed[64] = { 0 };
struct timeval timeout;
buffer = malloc(IPERF_BUFSZ);
if (buffer == NULL)
{
return;
}
sock = socket(PF_INET, SOCK_DGRAM, 0);
if(sock < 0)
{
rt_kprintf("can't create socket!! exit\n");
return;
}
server.sin_family = PF_INET;
server.sin_port = htons(param.port);
server.sin_addr.s_addr = inet_addr("0.0.0.0");
timeout.tv_sec = 2;
timeout.tv_usec = 0;
if (setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO, &timeout, sizeof(timeout)) == -1)
{
rt_kprintf("setsockopt failed!!");
closesocket(sock);
free(buffer);
return;
}
if (bind(sock, (struct sockaddr *)&server, sizeof(struct sockaddr_in)) < 0)
{
rt_kprintf("iperf server bind failed!! exit\n");
closesocket(sock);
free(buffer);
return;
}
while (param.mode != IPERF_MODE_STOP)
{
tick1 = rt_tick_get();
tick2 = tick1;
lost = 0;
total = 0;
sentlen = 0;
while ((tick2 - tick1) < (RT_TICK_PER_SECOND * 5))
{
r_size = recvfrom(sock, buffer, IPERF_BUFSZ, 0, (struct sockaddr *)&sender, (socklen_t*)&sender_len);
if (r_size > 12)
{
pcount = ntohl(buffer[0]);
if (last_pcount < pcount)
{
lost += pcount - last_pcount - 1;
total += pcount - last_pcount;
}
else
{
last_pcount = pcount;
}
last_pcount = pcount;
sentlen += r_size;
}
tick2 = rt_tick_get();
}
if (sentlen > 0)
{
f = (float)(sentlen * RT_TICK_PER_SECOND / 125 / (tick2 - tick1));
f /= 1000.0f;
snprintf(speed, sizeof(speed), "%.4f Mbps! lost:%d total:%d\n", f, lost, total);
rt_kprintf("%s", speed);
}
}
free(buffer);
closesocket(sock);
}
static void iperf_client(void *thread_param)
{
int i;
int sock;
int ret;
int tips = 1;
uint8_t *send_buf;
rt_uint64_t sentlen;
rt_tick_t tick1, tick2;
struct sockaddr_in addr;
char speed[32] = { 0 };
send_buf = (uint8_t *) malloc(IPERF_BUFSZ);
if (!send_buf) return ;
for (i = 0; i < IPERF_BUFSZ; i ++)
send_buf[i] = i & 0xff;
while (param.mode != IPERF_MODE_STOP)
{
sock = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
if (sock < 0)
{
rt_kprintf("create socket failed!\n");
rt_thread_delay(RT_TICK_PER_SECOND);
continue;
}
addr.sin_family = PF_INET;
addr.sin_port = htons(param.port);
addr.sin_addr.s_addr = inet_addr((char *)param.host);
ret = connect(sock, (const struct sockaddr *)&addr, sizeof(addr));
if (ret == -1)
{
if (tips)
{
rt_kprintf("Connect to iperf server faile, Waiting for the server to open!\n");
tips = 0;
}
closesocket(sock);
rt_thread_delay(RT_TICK_PER_SECOND);
continue;
}
rt_kprintf("Connect to iperf server successful!\n");
{
int flag = 1;
setsockopt(sock,
IPPROTO_TCP, /* set option at TCP level */
TCP_NODELAY, /* name of option */
(void *) &flag, /* the cast is historical cruft */
sizeof(int)); /* length of option value */
}
sentlen = 0;
tick1 = rt_tick_get();
while (param.mode != IPERF_MODE_STOP)
{
tick2 = rt_tick_get();
if (tick2 - tick1 >= RT_TICK_PER_SECOND * 5)
{
float f;
f = (float)(sentlen * RT_TICK_PER_SECOND / 125 / (tick2 - tick1));
f /= 1000.0f;
snprintf(speed, sizeof(speed), "%.4f Mbps!\n", f);
rt_kprintf("%s", speed);
tick1 = tick2;
sentlen = 0;
}
ret = send(sock, send_buf, IPERF_BUFSZ, 0);
if (ret > 0)
{
sentlen += ret;
}
if (ret < 0) break;
}
closesocket(sock);
rt_thread_delay(RT_TICK_PER_SECOND * 2);
rt_kprintf("Disconnected, iperf server shut down!\n");
tips = 1;
}
free(send_buf);
}
void iperf_server(void *thread_param)
{
uint8_t *recv_data;
socklen_t sin_size;
rt_tick_t tick1, tick2;
int sock = -1, connected, bytes_received;
rt_uint64_t recvlen;
struct sockaddr_in server_addr, client_addr;
char speed[32] = { 0 };
fd_set readset;
struct timeval timeout;
recv_data = (uint8_t *)malloc(IPERF_BUFSZ);
if (recv_data == RT_NULL)
{
rt_kprintf("No memory\n");
goto __exit;
}
sock = socket(AF_INET, SOCK_STREAM, 0);
if (sock < 0)
{
rt_kprintf("Socket error\n");
goto __exit;
}
server_addr.sin_family = AF_INET;
server_addr.sin_port = htons(param.port);
server_addr.sin_addr.s_addr = INADDR_ANY;
memset(&(server_addr.sin_zero), 0x0, sizeof(server_addr.sin_zero));
if (bind(sock, (struct sockaddr *)&server_addr, sizeof(struct sockaddr)) == -1)
{
rt_kprintf("Unable to bind\n");
goto __exit;
}
if (listen(sock, 5) == -1)
{
rt_kprintf("Listen error\n");
goto __exit;
}
timeout.tv_sec = 3;
timeout.tv_usec = 0;
while (param.mode != IPERF_MODE_STOP)
{
FD_ZERO(&readset);
FD_SET(sock, &readset);
if (select(sock + 1, &readset, RT_NULL, RT_NULL, &timeout) == 0)
continue;
sin_size = sizeof(struct sockaddr_in);
connected = accept(sock, (struct sockaddr *)&client_addr, &sin_size);
rt_kprintf("new client connected from (%s, %d)\n",
inet_ntoa(client_addr.sin_addr), ntohs(client_addr.sin_port));
{
int flag = 1;
setsockopt(connected,
IPPROTO_TCP, /* set option at TCP level */
TCP_NODELAY, /* name of option */
(void *) &flag, /* the cast is historical cruft */
sizeof(int)); /* length of option value */
}
recvlen = 0;
tick1 = rt_tick_get();
while (param.mode != IPERF_MODE_STOP)
{
bytes_received = recv(connected, recv_data, IPERF_BUFSZ, 0);
if (bytes_received <= 0) break;
recvlen += bytes_received;
tick2 = rt_tick_get();
if (tick2 - tick1 >= RT_TICK_PER_SECOND * 5)
{
float f;
f = (float)(recvlen * RT_TICK_PER_SECOND / 125 / (tick2 - tick1));
f /= 1000.0f;
snprintf(speed, sizeof(speed), "%.4f Mbps!\n", f);
rt_kprintf("%s", speed);
tick1 = tick2;
recvlen = 0;
}
}
rt_kprintf("client disconnected (%s, %d)\n",
inet_ntoa(client_addr.sin_addr), ntohs(client_addr.sin_port));
if (connected >= 0) closesocket(connected);
connected = -1;
}
__exit:
if (sock >= 0) closesocket(sock);
if (recv_data) free(recv_data);
}
void iperf_usage(void)
{
rt_kprintf("Usage: iperf [-s|-c host] [options]\n");
rt_kprintf(" iperf [-h|--stop]\n");
rt_kprintf("\n");
rt_kprintf("Client/Server:\n");
rt_kprintf(" -p # server port to listen on/connect to\n");
rt_kprintf("\n");
rt_kprintf("Server specific:\n");
rt_kprintf(" -s run in server mode\n");
rt_kprintf("\n");
rt_kprintf("Client specific:\n");
rt_kprintf(" -c <host> run in client mode, connecting to <host>\n");
rt_kprintf("\n");
rt_kprintf("Miscellaneous:\n");
rt_kprintf(" -h print this message and quit\n");
rt_kprintf(" --stop stop iperf program\n");
rt_kprintf(" -u testing UDP protocol");
return;
}
int iperf(int argc, char **argv)
{
int mode = 0; /* server mode */
char *host = NULL;
int port = IPERF_PORT;
int use_udp = 0;
int index = 1;
if (argc == 1)
{
goto __usage;
}
if (strcmp(argv[1], "-u") == 0)
{
index = 2;
use_udp = 1;
}
if (strcmp(argv[index], "-h") == 0) goto __usage;
else if (strcmp(argv[index], "--stop") == 0)
{
/* stop iperf */
param.mode = IPERF_MODE_STOP;
return 0;
}
else if (strcmp(argv[index], "-s") == 0)
{
mode = IPERF_MODE_SERVER; /* server mode */
/* iperf -s -p 5000 */
if ((argc == 4) || (argc == 5))
{
if (strcmp(argv[index + 1], "-p") == 0)
{
port = atoi(argv[index + 2]);
}
else goto __usage;
}
}
else if (strcmp(argv[index], "-c") == 0)
{
mode = IPERF_MODE_CLIENT; /* client mode */
if (argc < 3) goto __usage;
host = argv[index + 1];
if ((argc == 5) || (argc == 6))
{
/* iperf -c host -p port */
if (strcmp(argv[index + 2], "-p") == 0)
{
port = atoi(argv[index + 3]);
}
else goto __usage;
}
}
else if (strcmp(argv[index], "-h") == 0)
{
goto __usage;
}
else goto __usage;
/* start iperf */
if (param.mode == IPERF_MODE_STOP)
{
rt_thread_t tid = RT_NULL;
param.mode = mode;
param.port = port;
if (param.host)
{
rt_free(param.host);
param.host = NULL;
}
if (host) param.host = rt_strdup(host);
if (use_udp)
{
if (mode == IPERF_MODE_CLIENT)
{
tid = rt_thread_create("iperfc", iperf_udp_client, RT_NULL,
2048, 20, 20);
}
else if (mode == IPERF_MODE_SERVER)
{
tid = rt_thread_create("iperfd", iperf_udp_server, RT_NULL,
2048, 10, 20);
}
}
else
{
if (mode == IPERF_MODE_CLIENT)
{
tid = rt_thread_create("iperfc", iperf_client, RT_NULL,
2048, 20, 20);
}
else if (mode == IPERF_MODE_SERVER)
{
tid = rt_thread_create("iperfd", iperf_server, RT_NULL,
2048, 20, 20);
}
}
if (tid) rt_thread_startup(tid);
}
else
{
rt_kprintf("Please stop iperf firstly, by:\n");
rt_kprintf("iperf --stop\n");
}
return 0;
__usage:
iperf_usage();
return 0;
}
#ifdef RT_USING_FINSH
#include <finsh.h>
MSH_CMD_EXPORT(iperf, the network bandwidth measurement tool);
#endif
#endif /* PKG_NETUTILS_IPERF */
|
Introduction {#sec1}
============
Owing to its hydrophobicity, chemical resistance, high temperature resistance, and excellent electrical properties, silicone rubber (SR) has been extensively used in various electrical applications, such as overhead transmission lines, power stations, and cable accessories.^[@ref1]−[@ref3]^ However, due to its organic nature, when suffering from dry-band arcing, a peculiar surface defect called tracking and erosion failure is unavoidable for silicone rubber, which leads to a great hidden danger for the security of the power system.^[@ref4]^ To improve the antitracking performance of SR, the addition of inorganic fillers, such as alumina trihydrate,^[@ref5]^ alumina (Al~2~O~3~),^[@ref6]^ silica (SiO~2~),^[@ref7]^ boron nitride,^[@ref8]^ and titanium dioxide,^[@ref9]^ is the most commonly used method. Unfortunately, a high loading (\>50 wt %) is required for the desired tracking and erosion resistance. In this case, the mechanical properties and processability of SR are inevitably seriously damaged.^[@ref10]^ Moreover, with the development of electrical engineering, a higher reliability of insulation materials is expected.^[@ref11]^ Therefore, it is imperative to develop a more efficient antitracking additive to enhance the antitracking performance of SR.
Introducing arc-quenching materials including melamine compounds, urea compounds, and guanine compounds into SR is considered to be one of the most promising methods to significantly enhance the tracking and erosion resistance of SR,^[@ref12]^ because they can rapidly evolve inert gases to quench the electric arcs through the deionizing and cooling effect during arcing.^[@ref13]^ Schmidt and co-workers^[@ref14]^ used 15 phr of melamine cyanurate (MC) in combination with 100 phr silica to enhance the tracking and erosion resistance of SR. The results showed that all samples containing MC passed the inclined plane (IP) test at 4.5 kV due to the arc-quenching ability of MC. However, melamine cyanurate also has poor compatibility with SR, resulting in deterioration of the mechanical properties. Our previous work^[@ref15]^ indicated that a small amount of ureido-containing siloxane (US) effectively enhanced the tracking and erosion resistance of addition-cure liquid silicone rubber (ALSR), and the correlation between the thermostability and antitracking properties was studied in detail. However, because dry-band electrical arc is essentially high temperature ionized gas,^[@ref16]^ during the arc discharge, ALSR was subjected to not only thermal attack but also plasma bombardment, resulting in degradation of the silicone chains. Moreover, our further studies found that many organic additives (e.g., polyborosiloxane, shown in [Figure S1](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf)) improved the thermostability of ALSR but had no effect on the tracking and erosion resistance of ALSR, clearly indicating that the suppressed mechanism was not solely dependent on the improvement of thermal stability. In addition, the solid residue formed during dry-band arcing also had a close relation with the tracking and erosion performance.^[@ref17]^ Therefore, an explanation based only on the improvement of thermal stability of ALSR/US seemed to be insufficient, and it is of scientific and industrial importance to better our understanding of the antitracking mechanism of ALSR/US.
To further study whether it is the peculiar characteristics of the ureido group that improve the tracking and erosion resistance of ALSR, a common nitrogen-containing silane, *N*-(β-aminoethyl)-γ-aminopropyltriethoxysilane (AEAPTES), which has the same nitrogen per molecule as that of the ureido-containing siloxane, was used to enhance the antitracking performance of ALSR. However, it is worth noting that AEAPTAS is also easily hydrolyzed when exposed to a wet environment, resulting in a decrease of its efficiency. As is well known, MQ silicone resin, consisting of a monofunctional chain element (R~3~SiO~1/2~, i.e., M) and tetrafunctional chain element (SiO~4/2~, i.e., Q), is widely used as a reinforcing filler for silicone rubber.^[@ref18],[@ref19]^ In addition, MQ silicone resin has excellent hydrolysis resistance, weathering ageing resistance, and radiation resistance.^[@ref20]−[@ref22]^ Therefore, if nitrogen-containing groups were attached to MQ silicone resin, there would be a strong possibility of endowing ALSR with superior tracking and erosion resistance and mechanical properties.
In this work, an amine-containing MQ silicone resin (A-MQ) was prepared by the hydrolytic condensation of tetraethoxysilane (TEOS), *N*-(β-aminoethyl)-γ-aminopropyltriethoxysilane, hexamethyldisiloxane (MM), and divinyltetramethyldisiloxane (M^Vi^M^Vi^). Then, A-MQ was introduced into addition-cure liquid silicone rubber to enhance the tracking and erosion resistance. The effect of A-MQ on the tracking and erosion resistance and mechanical properties of ALSR was investigated. The possible antitracking mechanism of A-MQ was further explored by laser Raman spectroscopy (LRS), thermogravimetry (TG), thermogravimetry-Fourier transform infrared spectroscopy (TG-FTIR), scanning electron microscopy (SEM), attenuated total reflection-Fourier transform infrared spectrometry (ATR-FTIR), and X-ray photoelectron spectroscopy (XPS).
Results and Discussion {#sec2}
======================
Mechanical Properties {#sec2.1}
---------------------
[Table [1](#tbl1){ref-type="other"}](#tbl1){ref-type="other"} shows the effect of A-MQ content on the mechanical properties of ALSR. As shown, the mechanical performance of ALSR showed an obvious improvement with the addition of A-MQ. The tensile strength, elongation at break, and tear strength of ALSR increased at first and then decreased with increasing A-MQ amount. When the content of A-MQ was 3 phr, the tensile strength and tear strength of ALSR reached their maximum with values of 8.9 and 36.7 kN·m^--1^, respectively. Because A-MQ has more vinyl groups, these can react with the Si--H groups to form a more concentrated crosslinking network in ALSR. The appropriate concentration of crosslinking network can effectively disperse the stress to more molecular chains, thus improving the mechanical properties.^[@ref25]^ However, when the crosslinking density is too high, the tensile strength and tear strength of ALSR decrease. When A-MQ was 1 phr, the elongation at break of ALSR reached its maximum with the value of 712%. The reason for this might be that A-MQ improves the dispersion of SiO~2~ in ALSR,^[@ref22]^ and the good dispersion of SiO~2~ suppresses the filler network to increase the elongation at break of ALSR.^[@ref26]^ When the content of A-MQ further increased, the crosslinking density of ALSR also increased and played the dominant role, resulting in the decrease of elongation at break of ALSR.
###### Effect of A-MQ Content on the Mechanical Properties of the ALSR Samples
content (phr) tensile strength (MPa) elongation at break (%) tear strength (kN·m^--1^) hardness (shore A)
--------------- ------------------------ ------------------------- --------------------------- --------------------
0 7.6 ± 0.3 638 ± 56 31.6 ± 4.1 41 ± 1
1 8.1 ± 0.2 712 ± 38 33.4 ± 3.6 42 ± 1
2 8.5 ± 0.4 698 ± 51 35.1 ± 3.2 43 ± 1
3 8.9 ± 0.2 703 ± 60 36.7 ± 5.3 44 ± 1
4 8.6 ± 0.3 678 ± 75 35.9 ± 3.0 44 ± 1
Inclined Plane Test {#sec2.2}
-------------------
[Figure [1](#fig1){ref-type="fig"}](#fig1){ref-type="fig"} presents the tracking and erosion performance of the ALSR samples with different A-MQ content. When the voltage of 4.5 kV was applied, ALSR without A-MQ failed quickly within 40 min due to the overcurrent. After A-MQ was added, the time to failure greatly increased. When A-MQ content was 2 phr or more, ALSR passed the IP test at 4.5 kV for all samples. In addition, A-MQ also decreased the eroded mass of ALSR. When the content of A-MQ was 4 phr, the eroded mass of ALSR/A-MQ was the least, decreasing by 67.8% compared to that of ALSR. As can be seen from [Figure S2](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf), in comparison with those of ALSR, the number of leakage current impulses of ALSR containing 4 phr A-MQ was greatly reduced, and the magnitude of leakage current also decreased. The results indicate that A-MQ not only suppressed the generation of arc discharge, but also reduced the intensity of arcs.
{#fig1}
Raman Spectroscopy Analysis {#sec2.3}
---------------------------
To investigate the effect of A-MQ on the tracking resistance of ALSR, LRS was adopted to characterize the residue of ALSR and ALSR/A-MQ (4 phr) after the IP test. [Figure [2](#fig2){ref-type="fig"}](#fig2){ref-type="fig"} shows the LRS spectra of the residue of ALSR and ALSR/A-MQ after the IP test. As can be seen in the spectrum of ALSR, the peaks at 1347 and 1595 cm^--1^ were assigned to the D band and G band, which are associated with the breathing mode of aromatic rings with dangling bonds in plane terminations and the bond stretching mode of the sp^2^ carbon pairs in both rings and chains, respectively.^[@ref27],[@ref28]^ Thus, the appearance of the D and G peaks indicated that graphitic carbon had been deposited on the surface of ALSR during the IP test. As reported,^[@ref17]^ the presence of carbon distorted the electric field distribution on the surface and increased the electrical field stress at the vicinity of the defected interface, resulting in severe damage of the surface (as shown in [Figure S3](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf)) and the decrease of tracking time. In the spectrum of ALSR/A-MQ, there were no peaks around 1200--1600 cm^--1^, indicating that no carbon was deposited on the surface during the IP test. This result implied that the incorporation of A-MQ suppressed the generation of carbon, which might account for the high tracking resistance of ALSR/A-MQ.
{#fig2}
Thermogravimetric Analysis {#sec2.4}
--------------------------
The continuous dry-band arc discharge releases a lot of heat, which leads to thermal oxidative degradation on the surface of ALSR.^[@ref29]^ Thus, it is necessary to investigate the effect of A-MQ on the thermal oxidative stability of ALSR. [Figure [3](#fig3){ref-type="fig"}](#fig3){ref-type="fig"} shows the TG and derivative thermogravimetry (DTG) curves of A-MQ, ALSR, ALSR/A-MQ (4 phr), and ALSR/A-MQ (cal.), and their characteristic data are summarized in [Table [2](#tbl2){ref-type="other"}](#tbl2){ref-type="other"}. The curve of ALSR/A-MQ (cal.) was calculated as shown in [eq [1](#eq1){ref-type="disp-formula"}](#eq1){ref-type="disp-formula"}where, *W*~1~ and *W*~2~ are the weight curves of A-MQ and ALSR, ω~1~ and ω~a~ are the contents of A-MQ and all of the ingredients in the formula; when A-MQ content is 4 phr, the ratio of ω~1~ to ω~a~ is 0.027. As can be seen, the *T*~5%~ of ALSR/A-MQ was almost the same as that of ALSR, but compared to ALSR/A-MQ (cal.), the *T*~5%~ of ALSR/A-MQ increased by 42.3 °C, indicating that A-MQ could improve the initial thermal stability of ALSR. In addition, in comparison with that of ALSR/A-MQ (cal.), the *R*~max~ of ALSR/A-MQ decreased by 23.0% and the residue at 800 °C increased by 18.2%. This result demonstrates that there were some chemical reactions occurring between A-MQ and the silicone chains during thermal degradation. [Figure S4](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf) shows the TG and DTG curves of ALSR, ALSR/Vi-MQ, ALSR/N-MQ, and ALSR/A-MQ, and their characteristic data are summarized in [Table S1](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf). As can be seen from [Figure S4 and Table S1](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf), the *T*~5%~ of ALSR/Vi-MQ and ALSR/N-MQ were almost the same as that of ALSR, but the residue at 800 °C increased by 6.7 and 11.6%, indicating that both Vi-MQ and N-MQ suppressed the degradation of ALSR. Interestingly, the increment of ALSR/A-MQ to ALSR (18.2%) was equal to the linear addition of those of ALSR/Vi-MQ and ALSR/N-MQ to ALSR (6.7 and 11.6%), revealing that A-MQ could suppress the degradation of ALSR due to the effect of vinyl and amino groups.
{#fig3}
###### Characteristic Data Obtained from TG Curves
sample *T*~5%~ (°C) *T*~max~ (°C) *R*~max~ (wt %·min^--1^) residue at 800 °C (wt %)
------------------ -------------- --------------- -------------------------- --------------------------
A-MQ 248.5 306.1 13.44 42.2
ALSR 441.1 566.2 7.21 51.6
ALSR/A-MQ 440.6 561.1 5.55 61.0
ALSR/A-MQ (cal.) 398.3 566.6 7.32 51.4
Evolved Gases Analysis {#sec2.5}
----------------------
To further study the effect of A-MQ on the thermal degradation of ALSR under an air atmosphere, TG-FTIR was used to analyze the evolved gas products. [Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"} shows the three-dimensional (3D) TG-FTIR spectra of the pyrolysis gases in the thermal degradation of (a) ALSR and (b) ALSR/4 phr A-MQ under an air atmosphere. The FTIR spectra of the total volatile products for various samples are shown in [Figure [5](#fig5){ref-type="fig"}](#fig5){ref-type="fig"}. For ALSR and ALSR/A-MQ, six small molecular gaseous species could be identified by their characteristic absorbance peaks: carbonyl compounds (1745 cm^--1^), cyclic oligomers (2966, 1264, 1074, 1026, and 849 cm^--1^), methane (3017 and 1304 cm^--1^), CO (2179 and 2114 cm^--1^), CO~2~ (2359 and 2314 cm^--1^), and H~2~O (3500--3700 cm^--1^).
{#fig4}
{#fig5}
[Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"} shows the evolution curves of the pyrolysis products as the FTIR absorbance of the pyrolysis products versus temperature, wherein the amount of gas released is reflected by the peak areas. As demonstrated, cyclic oligomers were the main evolved gas for both ALSR and ALSR/A-MQ. Moreover, it has been reported that cyclic oligomers can activate discharges produced on the surface of poly(dimethylsiloxane) (PDMS), which promote stable dry-band arcing and increase the intensity of arcing.^[@ref30]^ Thus, the variation of evolution of the cyclic oligomers versus temperature attracted the most concern. As shown in [Figure [7](#fig7){ref-type="fig"}](#fig7){ref-type="fig"}, the release of cyclic oligomers was induced by an unzipping reaction and random scission.^[@ref31]^ The unzipping reaction was triggered by the silanol groups, which were generated by oxidation of the side groups, along with CH~2~O. As can be seen from [Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"}a,c, the evolution of CH~2~O and cyclic oligomers from ALSR/A-MQ was the same as that from ALSR between 342 and 468 °C, indicating that at the early stage, cyclic oligomers were evolved mainly by an unzipping reaction, and A-MQ did not suppress the initial oxidation of the side groups. The release of CH~4~ occurred via the cleavage of methyl groups by a radical mechanism, which promoted the formation of a tight silicone network (shown in [Figure [7](#fig7){ref-type="fig"}](#fig7){ref-type="fig"}d).^[@ref32]^ From [Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"}b,c it can be seen that the amount of evolved CH~4~ from ALSR/A-MQ was much more than that from ALSR, and the evolution of cyclic oligomers from ALSR/A-MQ dramatically decreased beyond 468 °C. To further clarify the effect of A-MQ, TG-FTIR tests of ALSR/Vi-MQ and ALSR/N-MQ were also conducted. As can be seen from [Figure S5b,c](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf), the evolution of CH~4~ from ALSR/Vi-MQ was almost the same as that from ALSR, but the evolution of cyclic oligomers from ALSR/Vi-MQ decreased compared with that from ALSR, indicating that the high number of vinyl groups in Vi-MQ improved the crosslinking density of ALSR, suppressing the generation of cyclic oligomers, but had little effect on the radical mechanism. For ALSR/N-MQ, the evolution of CH~4~ increased and cyclic oligomers decreased, compared with ALSR, revealing that the amino groups in N-MQ enhanced the catalytic effect of Pt on the radical mechanism. Therefore, the effect of A-MQ was attributed to two aspects. On the one hand, the amino groups in A-MQ significantly enhanced the catalytic effect of Pt on the radical mechanism, which promoted the formation of the compact and thermostable ceramic layer.^[@ref33]^ Such a ceramic layer could protect silicone chains from further degradation. On the other hand, the high vinyl group content in A-MQ increased the crosslinking density in ALSR, which restricted the movement of silicone chains and improved the activation energy of the occurrence of random scission.^[@ref34]^ At elevated temperature, Si--CH~3~ in polysiloxane molecules and CH~4~ are very likely to transform into carbon.^[@ref35]^ As shown in [Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"}a,d--f, the evolution of CH~2~O, CO, CO~2~, and H~2~O from ALSR/A-MQ increased above 557 °C, in comparison with that from ALSR. As is clear from [Figure S5a,d--f](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf), when Vi-MQ was added, the evolution of CH~2~O, CO, CO~2~, and H~2~O from ALSR decreased. However, the evolution of CH~2~O, CO, CO~2~, and H~2~O from ALSR/N-MQ increased, compared with that from ALSR. The results indicate that at elevated temperature, amino groups in A-MQ promote the oxidation of Si--CH~3~ and CH~4~, reducing the amount of carbon deposited on the surface of ALSR.
{#fig6}
{#fig7}
The arc discharge can generate plasma, which produces electron, positive and negative ion, and free radical bombardments on the surface of silicone rubber.^[@ref36]^ These bombardments result in the thermal oxidation and bond scission of polysiloxane molecules, forming surface char residue.^[@ref37]^ This residue increases the intensity of the dry-band arc, making silicone rubber tracking easily, thus, plasma bombardment is also a main cause of tracking and erosion of SR. To further investigate the effect of A-MQ on the tracking resistance, a low temperature plasma jet device was used to treat ALSR and ALSR/A-MQ (4 phr).
Morphology Analysis {#sec2.6}
-------------------
[Figure [8](#fig8){ref-type="fig"}](#fig8){ref-type="fig"} shows the SEM images of ALSR and ALSR/A-MQ after plasma treatment for different lengths of time. As can be seen, there was a noticeable difference in the surfaces of ALSR and ALSR/A-MQ after plasma treatment. For ALSR, the local surface was seriously damaged, the polysiloxane was eroded, and loosely bound filler appeared on the surface after 1 h of plasma treatment. With increasing treatment time, the erosion became more and more serious, and the damage area became increasingly extended. As for ALSR/A-MQ, the surface was smooth, and there were no noticeable defects after 1 h of treatment, indicating that A-MQ suppressed the bombardments, protecting the polysiloxane molecules from degradation. When the treatment time was 2 h, there was no erosion of polysiloxane or precipitation of filler particles on the surface, but some cracks appeared. When the treatment time reached 4 h, the amount of cracks increased greatly. Surprisingly, there were still no erosion defects.
{#fig8}
ATR-FTIR Analysis {#sec2.7}
-----------------
[Figure [9](#fig9){ref-type="fig"}](#fig9){ref-type="fig"} shows the ATR-FTIR spectra of ALSR and ALSR/A-MQ after plasma treatment for different lengths of time. The polysiloxane molecule is mainly composed of Si--CH~3~ and Si--O structures, and the corresponding characteristic peaks are located at 1260 and 1020 cm^--1^.^[@ref38]^ As can be seen from [Figure [10](#fig10){ref-type="fig"}](#fig10){ref-type="fig"}, with increasing treatment time, for ALSR and ALSR/A-MQ, the peak intensities of Si--CH~3~ and Si--O both continuously decreased, indicating that plasma could destroy the side groups and backbone of polysiloxane molecules.
{#fig9}
{#fig10}
To acquire more information about the chemical changes of the species, the absorption peak ratio of Si--CH~3~ to Si--O was calculated. The ratio of the absorption peaks (Si--CH~3~ to Si--O) of the untreated sample was defined as 100%. A reduction in the absorption ratio shows a degree of surface deterioration.^[@ref39]^ The variations of peak ratio (Si--CH~3~ to Si--O) of ALSR and ALSR/A-MQ with treatment time are shown in [Figure [10](#fig10){ref-type="fig"}](#fig10){ref-type="fig"}. As can be seen, with increasing treatment time, the ratios of ALSR and ALSR/A-MQ both decreased; when the treatment time reached 3 h, the ratios of ALSR and ALSR/A-MQ tended to remain constant. However, over the whole treatment time, the ratio of ALSR/A-MQ was much higher than that of ALSR. The results further confirm that A-MQ protected the polysiloxane molecule from plasma radiation.
XPS Analysis {#sec2.8}
------------
[Figure [11](#fig11){ref-type="fig"}](#fig11){ref-type="fig"} shows the C 1s XPS spectra of the surface of ALSR and ALSR/A-MQ after 4 h of plasma treatment, and the relevant characteristic parameters are summarized in [Table [3](#tbl3){ref-type="other"}](#tbl3){ref-type="other"}. After 4 h of plasma treatment, the C 1s spectrum of the surface of ALSR was split into five peaks. The peak at 284.5 eV was assigned to Si--C/C--H in the silicone chains.^[@ref40]^ The peaks at around 288.1 and 289.3 eV were attributed to C=O and O--C=O, respectively,^[@ref41]^ which were formed by the oxidation of Si--CH~3~ in the silicone chains. The peak at 285.5 eV was assigned to C--Si in the cross-linked network.^[@ref42]^ The peak at 284.8 eV was assigned to C=C in the aromatic species,^[@ref43]^ indicating that plasma promoted cleavage of the side groups of polysiloxane molecules to generate graphitized carbons. When A-MQ was added, with the same duration of plasma treatment, the structural content of C=O and O--C=O decreased, and the content of Si--C/C--H and C--Si in the crosslinking increased by 15.8 and 200%, respectively. This result can be explained by the Pt-catalyzed effects, whereby C--H of the methylene group adjacent to the amine group in A-MQ was activated to seize the polysiloxane macroradicals,^[@ref44]^ which formed a tight crosslinking network (as shown in [Figure [12](#fig12){ref-type="fig"}](#fig12){ref-type="fig"}). In addition, the increase of crosslinking density hindered the movement of macroradical chain segments to suppress further degradation of silicone chains. Furthermore, it is noteworthy that the peak of graphitized carbon disappeared in the C 1s spectrum of the surface of ALSR/A-MQ, indicating that A-MQ also suppressed the formation of graphitized carbon, which is in accordance with the Raman results.
{#fig11}
{#fig12}
###### Characteristic Parameters of XPS Spectra for ALSR and ALSR/A-MQ
area (%)
----------------------- ------- ---------- ------
Si--CH~3~ 284.5 68.5 79.3
graphite 284.8 22.3 0
Si--C in crosslinking 285.5 6.3 18.9
C=O 288.1 1.7 0.7
C=O--O 289.3 1.2 1.1
Conclusions {#sec3}
===========
Amine-containing MQ silicone resin was successfully synthesized by the hydrolytic condensation of TEOS, AEAPTES, MM, and M^Vi^M^Vi^. A-MQ could effectively enhance the tracking and erosion resistance of ALSR. When 4 phr of A-MQ was added, all test samples passed the inclined plane test at 4.5 kV, and the erosion mass decreased from 2.86 to 0.92 g. In addition, the mechanical properties were also enhanced. The LRS results revealed that A-MQ suppressed the generation of carbon during the arc discharge. The TG and TG-FTIR results indicated that at elevated temperature, A-MQ promoted crosslinking of the polysiloxane molecules and suppressed the generation of cyclic oligomers, which reduced the intensity of the electrical arc. The SEM, ATR-FTIR, and XPS results revealed that when suffering from plasma bombardment, which was produced by arc discharge, A-MQ could protect the silicone chains from degradation and eliminated the carbon deposited on the surface.
Experimental Section {#sec4}
====================
Materials {#sec4.1}
---------
Tetraethoxysilane, hydrochloric acid (HCl, 36 wt %), and toluene were obtained from Guangzhou Chemical Reagent Co., Ltd., China. Hexamethyldisiloxane and divinyltetramethyldisiloxane were purchased from Shanghai Jiancheng Industrial Co., Ltd., China. *N*-(β-Aminoethyl)-γ-aminopropyltriethoxysilane was provided by Xiya Regent Co., Ltd., China. Anhydrous magnesium sulfate (MgSO~4~) was supplied by Sinopharm Chemical Reagent Co., Ltd., China. Anhydrous ethanol (EtOH) and anhydrous sodium bicarbonate (NaHCO~3~) were purchased from Tianjin Fuchen Chemical Reagent Co., Ltd., China. Vinyl-terminated poly(dimethylsiloxane) (VPDMS, viscosity: 24 320 mPa**·**s and vinyl content: 0.28 mol %) was supplied by Maigao Hightech Materials Co., Ltd., China. Poly(hydromethylsiloxane) (PHMS, viscosity: 160 mPa**·**s and hydride content: 0.50 wt %), platinum(0)-1,3-divinyl-1,1,3,3-tetramethydisiloxane complex (Karstedt's catalyst), and 1-ethynylcyclohexanol (inhibitor) were purchased from Guangzhou Xiyou New Material Technology Co., Ltd., China. Fumed silica possessing a specific surface area of 200 m^2^·g^--1^ was supplied by Tokuyama traces, Japan.
Preparation of A-MQ {#sec4.2}
-------------------
In a 250 mL four-neck flask, 17.8 g of MM and 2.2 g of M^Vi^M^Vi^ were added to a solution composed of 10.8 g of HCl, 10.0 g of EtOH, and 14.4 g of deionized water, and the reaction was heated at 70 °C for 30 min with stirring. Then, 41.6 g of TEOS was added dropwise into the flask for 3 h, followed by stirring for an additional 30 min. Subsequently, 1.2 g of AEAPTES was added dropwise to the solution under stirring for 1 h at 70 °C. After the reaction was finished, 80 mL of toluene was added to the solution and mixed well. Then, the organic layer was separated, neutralized with NaHCO~3~, dried with MgSO~4~, and filtered. By removing the solvent under vacuum, A-MQ was obtained as a faint yellow viscous liquid. The vinyl group content and nitrogen content in A-MQ was 2.17 and 0.74 wt % respectively, which was determined by iodometric titration^[@ref23]^ and hydrochloric titration.^[@ref24]^ The synthetic illustration of A-MQ is shown in [Figure [13](#fig13){ref-type="fig"}](#fig13){ref-type="fig"}, and the chemical structure of A-MQ was determined by FTIR, ^1^H NMR, and gel permeation chromatography (GPC), as shown in [Figures S6 and S7](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf): FTIR (KBr, cm^--1^): 3400 (νN--H), 3052 (νC--H in Si--CH=CH~2~), 2889--2974 (νC--H in CH~2~ and CH~3~), 1530 (νC--N), 1250, 750 (vSi--CH~3~), 980--1200 (νSi--O--Si); ^1^H NMR (600 MHz, CDCl~3~, δ, ppm): 5.95 (t, −C[H]{.ul}=C[H]{.ul}~2~), 2.35 (t, −NH~2~C[H]{.ul}~2~CH~2~NH−), 2.22 (t, −NH~2~CH~2~C[H]{.ul}~2~NH−), 2.17 (s, −NHC[H]{.ul}~2~CH~2~−), 1.60 (s, −NHCH~2~C[H]{.ul}~2~CH~2~−), 1.22 (m, −N[H]{.ul}~2~CH~2~CH~2~N[H]{.ul}−), 0.85 (m, −NHCH~2~CH~2~C[H]{.ul}~2~−), 0--0.2 (m, Si--CH~3~); GPC: *M*~n~ = 1010, *M*~w~/*M*~n~ = 1.1.
{#fig13}
Preparation of ALSR Samples with Different A-MQ Content {#sec4.3}
-------------------------------------------------------
VPDMS and fume silica were mixed well by a kneader to obtain masterbatch. Subsequently, masterbatch, PHMS, A-MQ, and 1-ethynylcyclohexanol were stirred vigorously. Then, Karstedt's catalyst was incorporated and mixed well. Finally, the mixture was vulcanized at 120 °C for 10 min under 8 MPa to obtain the ALSR sample. The formula of ALSR is listed in [Table [4](#tbl4){ref-type="other"}](#tbl4){ref-type="other"}.
###### Formula of the ALSR Samples
component content (phr)
----------------------- ---------------------------
VPDMS 100
PHMS *n*~SiH~/*n*~vinyl~ = 1.7
SiO~2~ 40
1-ethynylcyclohexanol 0.06
A-MQ 0--4
Karstedt's catalyst 0.38
Characterization {#sec4.4}
----------------
FTIR and ATR-FTIR spectra of the samples were obtained using a Bruker Tensor 27 spectrometer over the wave number range of 400--4000 cm^--1^. The liquid samples were coated on the surface of KBr tablets.
^1^H NMR spectra were obtained by using a Bruker Avance III HD 600 NMR spectrometer with CDCl~3~ as the solvent and tetramethylsilane as the internal standard.
Gel permeation chromatography (GPC) was performed using a Waters 515 HPLC pump (Waters) equipped with a Shodex K-G guard column and a Shodex K-804L chromatographic column. Detection was achieved using a Waters 2414 refraction index detector, and the sample was analyzed at 30 °C using chloroform as the eluent at a flow rate of 1 mL·min^--1^. The instrument was calibrated using narrow polydispersity polystyrene standards.
Tensile and tear tests of the cured samples were conducted on a universal testing machine (UT-1080, China) according to ASTM D 412 and ASTM D 624, respectively. The shore A hardness was measured with a Shore A durometer (LX-A, Shanghai Yuanling Instruments Factory, China) according to ASTM D 2240.
Tracking and erosion property analysis was carried out by an inclined plane tracking and erosion resistance test apparatus (DX8427, Dongguan Daxian Instruments Co., Ltd., China) according to IEC 60587-2003 standard. A schematic diagram of the inclined plane test setup is shown in [Figure [14](#fig14){ref-type="fig"}](#fig14){ref-type="fig"}, and digital photos of the tracking equipment and sample setup are shown in [Figure S8](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf). Each test sample was 120 × 50 × 6 mm^3^ and mounted at an inclination of 45°. Two electrodes were fixed on the surface of each test sample with a distance of 50 mm. During the test, a constant alternating current voltage of 4.5 kV was applied to each sample, along with a flow rate of 0.6 mL·min^--1^ of standardized conductive solution (0.10 wt % NH~4~Cl and 0.02 wt % isooctylphenoxypolyethoxyethanol). Five specimens were tested for each formulation. When the leakage current exceeded 60 mA for 2 s, the test apparatus recognized this moment as the time to failure. After 6 h of IP testing, a sample without excess current was regarded as having passed. After the IP test, the eroded portion of the test samples was cleared away, and the decreased mass of the specimen was recorded as the eroded mass.
{#fig14}
Laser Raman spectroscopy of the residue of SR after the IP test was determined by a Raman microspectrometer (Renishaw inVia, Renishaw Co., Britain) at an optical range from 3000 to 100 cm^--1^ with a 532 nm helium--neon laser source.
Thermogravimetric analysis was carried out by using a thermogravimeter (TG209, Netzsch Instruments Co., Germany) from 30 to 900 °C at a linear heating rate of 20 °C·min^--1^ under an air atmosphere. The samples were measured in an alumina crucible with a weight of 5--10 mg.
The TG-FTIR instrument consists of a thermogravimeter (TG209, Netzsch Instruments Co., Germany), a Fourier transform infrared spectrometer (Tensor 27, Bruker Optics Inc., Germany), and a transfer tube with an inner diameter of 1 mm connecting the TG and the infrared cell. The investigation was carried out from 30 to 900 °C at a linear heating rate of 20 °C·min^--1^ under an air flow of 30 mL·min^--1^. To reduce the possibility of pyrolysis gas condensing along the transfer tube, the temperatures of the infrared cell and transfer tube were set to 230 °C.
A low temperature plasma jet device (PlasmaFlecto 10, Plasmatechnology GmbH Co., Germany) was used to treat the ALSR samples. Each treatment sample was 10 × 10 × 2 mm^3^. The test was conducted under an air atmosphere with a power of 300 W. The highest temperature, which mainly appeared at the discharge area, was no more than 60 °C. Before treatment, all samples were cleaned with isopropanol and deionized water, separately.
The morphology of the surface of the SR samples after plasma treatment was investigated via field-emission scanning electron microscopy (Merlin Carl, Zeiss Jena, Co, Germany) at an acceleration voltage of 5 kV. Prior to measuring, samples were coated with a thin gold layer by means of a vacuum sputter to improve electrical conductivity.
The attenuated total reflection (ATR) technique enables identification of specific molecules and groups located in the surface layer, typically 1--10 μm deep. In this paper, attenuated total reflection-Fourier transform infrared spectroscopy (Bruker Tensor 27) was used to study the chemical structure of the SR surface after plasma treatment.
X-ray photoelectron spectroscopy was recorded on a Kratos Axis Ultra DLD X-ray photoelectron spectrometer by employing a monochromatic Al Kα X-ray source.
The Supporting Information is available free of charge on the [ACS Publications website](http://pubs.acs.org) at DOI: [10.1021/acsomega.7b00904](http://pubs.acs.org/doi/abs/10.1021/acsomega.7b00904).Tracking and erosion resistance and thermal stability of ALSR/PBS samples, leakage current of ALSR and ALSR/A-MQ (4 phr) during the IP test, digital photos of ALSR and ALSR/A-MQ (4 phr) after the IP test at 4.5 kV, TG and TG-FTIR results of ALSR samples in air, FTIR and ^1^H NMR spectra of A-MQ, GPC curve of A-MQ, and digital photos of tracking equipment and sample setup ([PDF](http://pubs.acs.org/doi/suppl/10.1021/acsomega.7b00904/suppl_file/ao7b00904_si_001.pdf))
Supplementary Material
======================
######
ao7b00904_si_001.pdf
The authors declare no competing financial interest.
The authors appreciate financial support from the National Science Foundation of China (Nos. 51573052 and 51403067) and the Pearl River S&T Nova Program of Guangzhou (201710010062).
|
// Copyright (c) 2019 Auriane Reverdell
//
// SPDX-License-Identifier: BSL-1.0
// Distributed under the Boost Software License, Version 1.0. (See accompanying
// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
#pragma once
#include <hpx/config.hpp>
#include <hpx/type_support/config/defines.hpp>
#include <hpx/modules/type_support.hpp>
#if HPX_TYPE_SUPPORT_HAVE_DEPRECATION_WARNINGS
#if defined(HPX_MSVC)
#pragma message("The header hpx/util/decay.hpp is deprecated, \
please include hpx/modules/type_support.hpp instead")
#else
#warning "The header hpx/util/decay.hpp is deprecated, \
please include hpx/modules/type_support.hpp instead"
#endif
#endif
|
Use of the pediatric intensive care unit for post-procedural monitoring in young children following microlaryngobronchoscopy: Impact on resource utilization and hospital cost.
To assess the frequency of post-procedural complications, medical interventions, and hospital costs associated with microlaryngobronchoscopy (MLB) in children prophylactically admitted for pediatric intensive care unit (PICU) monitoring for age ≤ 2 years. We performed a single-center, retrospective, descriptive study within a 44-bed PICU in a stand-alone, tertiary, pediatric referral center. Inclusion criteria were age ≤2 years and pre-procedural selection of prophylactic PICU monitoring after MLB between January 2010 and December 2015. Children were excluded for existing tracheostomy, if undergoing concurrent non-otolaryngeal procedures, or if intubated at the time of PICU admission. Primary outcomes were the development of major and minor procedural complications and medical rescue interventions. Secondary outcomes were hospital cost and length of stay (LOS). One hundred and eight subjects met inclusion criteria with a median age of 5.3 (IQR: 2.6-10.9) months. A majority (86%) underwent therapeutic instrumentation in addition to diagnostic MLB. There were no observed major complications or rescue interventions. Minor complications were noted within 5 h of monitoring and included isolated stridor (24%), desaturation <90% (10%), and nausea/emesis (8%). Minor interventions included supplemental oxygen via regular nasal cannula (39%), single-dose inhaled racemic epinephrine (19%), single-dose systemic corticosteroids (19%), or high flow nasal cannula (HFNC) therapy (4%). Save for two cases of HFNC, interventions were completed or discontinued within 5 h. Median PICU LOS was 1.1 days and median cost was $9650 (IQR: $8235- $14,861) per encounter. Estimated cost of same day observation in our post anesthesia care unit (PACU) following MLB without PICU admission is $1921 per encounter. In children ≤ 2 years of age prophylactically admitted for PICU observation, we did not observe severe complications or major interventions after MLB. Minor interventions and complications were noted early during post-procedural monitoring. PICU monitoring was substantially more expensive than same-day PACU observation. Young age as the sole criteria for prophylactic PICU monitoring after diagnostic or therapeutic MLB may be unjustified when comparable, cost-conscious care can be achieved in a PACU setting. Prior to pre-procedural selection of PICU monitoring, we recommend a broad contextual risk assessment including a review of comorbidities, operative plan, and intended anesthetic exposure. |
---
abstract: |
Consider a system of particles moving independently as Brownian motions until two of them meet, when the colliding pair annihilates instantly. The construction of such a system of annihilating Brownian motions (aBMs) is straightforward as long as we start with a finite number of particles, but is more involved for infinitely many particles. In particular, if we let the set of starting points become increasingly dense in the real line it is not obvious whether the resulting systems of aBMs converge and what the possible limit points (entrance laws) are. In this paper, we show that aBMs arise as the interface model of the continuous-space voter model. This link allows us to provide a full classification of entrance laws for aBMs. We also give some examples showing how different entrance laws can be obtained via finite approximations. Further, we discuss the relation of the continuous-space voter model to the stepping stone and other related models. Finally, we obtain an expression for the $n$-point densities of aBMs starting from an arbitrary entrance law.
*2010 Mathematics Subject Classification*: Primary 60K35, Secondary 60J68, 60H15.
bibliography:
- 'duality.bib'
---
[**Entrance laws for annihilating Brownian motions\
and the continuous-space voter model** ]{}\
<span style="font-variant:small-caps;">Matthias Hammer[^1], Marcel Ortgiese[^2] and Florian Völlering[^3]</span>\
[***Keywords.***]{} Annihilating Brownian motions, entrance laws, voter model, stepping stone model, symbiotic branching, moment duality.
Introduction {#intro}
============
Consider a system of particles moving independently as Brownian motions such that whenever two of them meet, the colliding pair annihilates instantly. As long as we start with a finite number of particles, the construction of such a system of annihilating Brownian motions (from now on called aBMs) is straightforward. It is also possible to start aBMs from infinitely many particles, provided that the initial positions do not accumulate, i.e. form a *discrete* and *closed* [(or equivalently, locally finite)]{} subset of the real line. The construction of such an infinite system is already not completely trivial, see e.g. [@TZ11 Sec. 4.1] or Section \[sec:construction\] below for some details. Thus a suitable state space for the evolution of aBMs is given by $${\mathcal{D}}:=\{{{\mathbf x}}\subseteq{\mathbb{R}}:{{\mathbf x}}\text{ is discrete and closed}\},$$ and for each ${{\mathbf x}}\in{\mathcal{D}}$ a system of aBMs starting from ${{\mathbf x}}$ can be constructed as a (strong) Markov process ${{\mathbf X}}^{{\mathbf x}}=({{\mathbf X}}^{{\mathbf x}}_t)_{t\ge0}$ taking values in ${\mathcal{D}}$.
Now let ${{\mathbf x}}_{n}\in{\mathcal{D}}$ be a sequence of discrete closed subsets of ${\mathbb{R}}$ which eventually become dense in the real line. We can ask the question, for which such sequences the corresponding aBM processes ${{\mathbf X}}^{{{\mathbf x}}_n}$ converge and what the possible limit points are. Intuitively, such a limit should correspond to a system of aBMs ‘started everywhere on the real line’. More formally, such a limit gives rise to an *entrance law* for the semigroup of aBMs on ${\mathcal{D}}$ (see below where we recall the formal definition). However, it is not clear a priori whether all asymptotically dense sets of starting points will lead to the same entrance law. This is in contrast to *coalescing* Brownian motions (from now on called cBMs), which have a monotonicity property. For cBMs, it is possible to add initial particles one by one, and as long as the asymptotic set of starting points is dense one ends up with a universal maximal object, the *Arratia flow*, see [@A79]. Thus in the coalescing case there is a unique maximal entrance law, where Brownian motions are started everywhere on the real line. When starting cBMs in all space-time points, the resulting object is called the *Brownian web*, see e.g. [@SSS17] for a recent survey.
For the annihilating case, in [@TZ11] Tribe and Zaboronski define a corresponding ‘maximal’ entrance law as a ‘thinned’ version of the maximal entrance law for cBMs (see Sec. 2.1 of their paper for the well-known thinning relation linking coalescing and annihilating systems). Moreover, they argue that this entrance law can be approximated by aBMs started from the lattice $\frac{1}{n}{\mathbb{Z}}$, or from points of a Poisson process with intensity $n$, by sending $n\to\infty$, but point out that the domain of attraction of this entrance law is not clear.
In this paper, we show that indeed different approximations of ${\mathbb{R}}$ by asymptotically dense sets will typically lead to different entrance laws for aBMs, as opposed to the case for cBMs. For example, if one starts a system of aBMs in ${{\mathbf x}}_{n}:=\frac1n{\ensuremath{\mathbb{Z}}}+\{0,\frac1{n^2}\}$ so that starting points appear in close-by pairs, then typically the pairs annihilate and in the limit there are no surviving annihilating Brownian motions at all. In our main result, Theorem \[thm:characterization\_entrance\_laws\], we will give a complete classification of entrance laws for aBMs via identification with measurable functions $u:{\ensuremath{\mathbb{R}}}\to[0,1]$.
Our classification of the entrance laws is based on a close connection of aBMs with the *continuous-space voter model*, which is a generalization of the classical discrete voter model to a continuous space setting. We will review this model and some of the relevant literature in Section \[sec:cSSM\].
As an application of this relation and the technique of duality, we can compute $n$-point densities for aBMs, i.e. the probability density of finding $n$ particles at given points. There has been some interest in these $n$-point densities recently, and indeed the main result in [@TZ11] is to show that a system of cBMs, but also of aBMs, started from the ‘maximal’ entrance law forms a Pfaffian point process and to give an expression for the densities. In particular, [@TZ11] show that these expressions can be used to derive large-time asymptotics.
In contrast, our result allows us to calculate $n$-point densities for any entrance law. For example, we can explicitly compute the $1$-particle density function and compare it to the density function under the entrance law constructed in [@TZ11]. We can show that the latter is only maximal when compared to homogeneous entrance laws, so that a more appropriate name would be ‘maximal homogeneous’. Our technique also gives an expression for $n$-point densities with $n >1$, which is however less explicit.
The paper is structured as follows: In Section \[sec:classification\], we will state [our results, namely the classification of entrance laws for aBMs and the corresponding $n$-point densities. ]{} In Section \[sec:cSSM\], we will explain the connection to the continuous-space voter model. We use the relation to the voter model and its duality in Section \[sec:proofs\] to prove the results of Section \[sec:classification\]. In the appendix, we recall in Section \[sec:construction\] how to construct aBMs starting from an infinite discrete closed set and prove [two technical results]{} in Section \[sec:technicalities\].
Notation and preliminaries {#ssec:notation}
--------------------------
The following notation and definitions will be used throughout the paper: Recall that ${\mathcal{D}}$ denotes the space of discrete closed subsets of ${\mathbb{R}}$,
and that we write ${{\mathbf x}}$ for a generic element of ${\mathcal{D}}$. With slight abuse of notation, we will occasionally use the same symbol for vectors and write also ${{\mathbf x}}=(x_1,\ldots,x_n)\in{\mathbb{R}}^n$. We denote by ${\ensuremath{\mathbb{R}}}^{n,\uparrow}:=\{{{\mathbf x}}\in{\ensuremath{\mathbb{R}}}^n: x_1<\cdots<x_n\}$ resp. ${\ensuremath{\mathbb{R}}}^{n,\downarrow}:=\{{{\mathbf x}}\in{\ensuremath{\mathbb{R}}}^n: x_1>\cdots>x_n\}$ the space of increasing resp. decreasing vectors in ${\ensuremath{\mathbb{R}}}^n$.
Moreover, recall that we denote by ${{\mathbf X}}^{{\mathbf x}}=({{\mathbf X}}_t^{{\mathbf x}})_{t\ge0}$ a countable system of annihilating Brownian motions starting from ${{\mathbf x}}\in{\mathcal{D}}$. See e.g. [@TZ11 Sec. 4.1] and Section \[sec:construction\] below for two possible approaches to the construction of ${{\mathbf X}}^{{\mathbf x}}$ in case that the initial condition ${{\mathbf x}}$ is countably infinite. Considering this system as a (strong) Markov process with state space ${\mathcal{D}}$, we write also ${{\mathbf X}}=({{\mathbf X}}_t)_{t\ge0}$ for the canonical process on the path space ${\mathcal{D}}^{[0,\infty)}$ and $({\mathbb{P}}_{{\mathbf x}})_{{{\mathbf x}}\in{\mathcal{D}}}$ for the corresponding family of probability measures such that ${{\mathbf X}}$ starts from ${{\mathbf x}}\in{\mathcal{D}}$ under ${\mathbb{P}}_{{\mathbf x}}$. The corresponding Markov semigroup on ${\mathcal{D}}$ will be denoted by $(P_t)_{t\ge0}$. Note that we did not mention any topology for ${\mathcal{D}}$. The right choice of topology is an important point which we will discuss in Section \[sec:topology\].
Finally, for ${{\mathbf x}}\in{\mathcal{D}}$ we denote by $$\label{eq:notation-cBMs}
{{\mathbf Y}}^{{\mathbf x}}_t=\{Y_t^{{\scriptscriptstyle{({x}})}} \,|\, x\in{{\mathbf x}}\},\qquad t\ge0$$ a system of *coalescing* Brownian motions starting from ${{\mathbf x}}$. Here, $Y_t^{{\scriptscriptstyle{({x}})}}$ is the position at time $t$ of the particle in the system which started in $x\in{{\mathbf x}}$ at time $t=0$. We may imagine that each particle follows the paths of a Brownian motion starting in $x$ until it collides with another motion, upon which the two particles are ‘merged’ and evolve together.
[Results]{} {#sec:classification}
===========
In this section we state our main results. We will classify entrance laws for aBMs by embedding ${\mathcal{D}}$ into a compact space and extending ${{\mathbf X}}$ to a Feller process on this space, for which all entrance laws are closable, and which we can describe explicitly.
Entrance laws
-------------
Recall that a family $\mu=(\mu_t)_{t>0}$ of probability measures on (the Borel $\sigma$-algebra of) ${\mathcal{D}}$ is called a *probability entrance law* for the semigroup $(P_t)_{t\ge0}$ if $$\label{eq:entrance_law}\mu_sP_{t-s}=\mu_t \qquad \text{for all }0<s<t.$$ See e.g. [@Li Appendix A.5] or [@Sharpe] for the general theory of entrance laws. Roughly speaking, an entrance law corresponds to a Markov process $({{\mathbf X}}_t)_{t>0}$ with time-parameter set $(0,\infty)$ and ‘without initial condition’, whose one-dimensional distributions are given by $\mu_t$.
Let $${\mathcal{M}}_1({\ensuremath{\mathbb{R}}}):=\{u(x)\,dx\,|\,u:{\ensuremath{\mathbb{R}}}\to[0,1]\text{ measurable}\}$$ denote the space of all absolutely continuous measures on ${\mathbb{R}}$ with densities taking values in $[0,1]$. We define an equivalence relation $\sim$ on ${\mathcal{M}}_1({\mathbb{R}})$ by identifying $u$ with $1-u$ and consider the quotient space $${\mathcal{V}}:= {\mathcal{M}}_1({\mathbb{R}}) /\! \sim.$$ We write $v=[u]=\{u,1-u\}$ for elements of ${\mathcal{V}}$, i.e. for the equivalence classes under $\sim$.
Our main result, Theorem \[thm:characterization\_entrance\_laws\], states that there is a bijective correspondence between probability entrance laws $(\mu_t)_{t>0}$ for the semigroup $(P_t)_{t\ge0}$ of aBMs on ${\mathcal{D}}$ and probability measures $\nu$ on ${\mathcal{V}}$. The subtle point is that this only works with the right topology on ${\mathcal{D}}$, which we will describe in the next subsection.
The topology on ${\mathcal{D}}$ {#sec:topology}
-------------------------------
In order to turn ${\mathcal{D}}$ into a topological space, as in [@TZ11] one may identify ${{\mathbf x}}\in{\mathcal{D}}$ with the locally finite point measure $\sum_{x\in{{\mathbf x}}}\delta_x$, thus embedding ${\mathcal{D}}$ into the space of locally finite measures, and use the topology of vague convergence. Note however that employing this topology leads to càdlàg but not continuous paths for the process ${{\mathbf X}}$, since at annihilation events the total mass of the finite point measure changes.
We will introduce a different (weaker) topology on ${\mathcal{D}}$ under which the paths of ${{\mathbf X}}$ are automatically continuous and which allows us to classify the entrance laws. The main idea is to regard the positions of the annihilating particles as ‘interfaces’ of two measures on the real line with complementary support, and to use these measures to obtain a topology better adapted to the evolution of aBMs. In order to make this precise, we need to introduce some additional notation and definitions. Recall that ${\mathcal{M}}_1({\ensuremath{\mathbb{R}}})$ denotes the space of all absolutely continuous measures on ${\mathbb{R}}$ with densities taking values in $[0,1]$. We will usually use the same symbol to denote the absolutely continuous measure and its density. We endow ${\mathcal{M}}_1({\mathbb{R}})$ with the vague topology, i.e. $u^{{\scriptscriptstyle{({n}})}}\to u$ in ${\mathcal{M}}_1({\ensuremath{\mathbb{R}}})$ iff $\langle u^{{\scriptscriptstyle{({n}})}},\phi\rangle\to\langle u,\phi\rangle$ for all $\phi\in{\mathcal{C}}_c({\ensuremath{\mathbb{R}}})$. It is easy to see that with this topology, ${\mathcal{M}}_1({\mathbb{R}})$ is a compact space, see Lemma \[lem:compact\] below. For $u\in{\mathcal{M}}_1({\mathbb{R}})$, we define the *interface* (of $u$ with its complement $1-u$) as $${\mathcal{I}}(u):={{\rm supp}}(u)\cap{{\rm supp}}(1-u),$$ where ${{\rm supp}}(u)$ denotes the measure-theoretic support of $u$, i.e. $${{\rm supp}}(u):=\{x\in{\mathbb{R}}\,|\, u\left(B_{\varepsilon}(x)\right)>0 \text{ for all }{\varepsilon}>0\} .$$ We call the elements of ${\mathcal{I}}(u)$ *interface points*. [Note that ${\mathcal{I}}(u)$ is always closed.]{} The subspace of all $u\in{\mathcal{M}}_1({\mathbb{R}})$ with discrete interface is denoted by $${\mathcal{M}}_1^d({\mathbb{R}}):=\{u\in{\mathcal{M}}_1({\mathbb{R}})\,|\,{\mathcal{I}}(u)\in{\mathcal{D}}\},$$ which is dense in ${\mathcal{M}}_1({\mathbb{R}})$, see Lemma \[lem:dense\] below. Note that for each $u\in{\mathcal{M}}_1^d({\mathbb{R}})$, we may choose a version of its density taking values in $\{0,1\}$ and which is locally constant on each of the countably many disjoint open intervals in ${\mathbb{R}}\setminus{\mathcal{I}}(u)$, where it takes the value $0$ or $1$ alternatingly. In particular, for $u\in{\mathcal{M}}_1^d({\mathbb{R}})$ the measure-theoretic and function-theoretic supports coincide.
When restricted to ${\mathcal{M}}_1^d({\mathbb{R}})$, the ‘interface operator’ gives us a mapping ${\mathcal{I}}:{\mathcal{M}}_1^d({\mathbb{R}})\to{\mathcal{D}}$ which is clearly surjective but not injective, since both $u$ and $1-u$ have the same interface. Thus with the equivalence relation $\sim$ on ${\mathcal{M}}_1({\mathbb{R}})$ identifying $u$ and $1-u$, we consider the quotient spaces $${\mathcal{V}}^d := {\mathcal{M}}_1^d({\mathbb{R}}) /\! \sim$$ and ${\mathcal{V}}= {\mathcal{M}}_1({\mathbb{R}}) /\!\!\sim$ introduced above. Endowed with the quotient topology, ${\mathcal{V}}$ is also compact and ${\mathcal{V}}^d$ is dense in ${\mathcal{V}}$. Note that the ‘interface operator’ ${\mathcal{I}}$ is well-defined on the equivalence classes and thus induces a mapping (which we denote by the same symbol) $${\mathcal{I}}:{\mathcal{V}}^d\to{\mathcal{D}},$$ which is easily seen to be a bijection and induces in a canonical way a topology on ${\mathcal{D}}$, generated by the system $$\{{\mathcal{I}}(U): U\subseteq{\mathcal{V}}^d\text{ open}\}.$$ By definition, this is the coarsest topology on ${\mathcal{D}}$ with respect to which ${\mathcal{I}}^{-1}:{\mathcal{D}}\to{\mathcal{V}}^d$ is continuous, and with this topology ${\mathcal{D}}$ is homeomorphic to ${\mathcal{V}}^d$. We note that this topology on ${\mathcal{D}}$ is strictly weaker than the topology used in [@TZ11].
[Classification of entrance laws]{}
-----------------------------------
Now we return to the aBM process $({{\mathbf X}}_t)_{t\ge0}$ on ${\mathcal{D}}$ with semigroup $(P_t)_{t\ge0}$. Via the homeomorphism ${\mathcal{I}}^{-1}$, it induces a semigroup $(T_t)_{t\ge0}$ on ${\mathcal{V}}^d$: $$\label{eq:defn_Q}
T_t(v;\cdot):=P_t({\mathcal{I}}(v);\cdot)\circ{\mathcal{I}},\qquad v\in{\mathcal{V}}^d,\;t\ge0.$$ Our main result states that this semigroup can be extended to a Feller semigroup $(\hat T_t)_{t\ge0}$ on the compact space ${\mathcal{V}}$ which can be used to characterize the entrance laws for aBMs.
In order to state this characterization precisely, we need the following notation:
Given $u\in{\mathcal{M}}_1({\mathbb{R}})$ and a system of cBMs $({{\mathbf Y}}_t^{{\mathbf x}})_{t\ge0}$ starting from some ${{\mathbf x}}=(x_1,\ldots,x_n)\in{\mathbb{R}}^n$, let $\left(\chi_y\right)_{y\in{{\mathbf Y}}_t^{{\mathbf x}}}$ be a family of Bernoulli random variables indexed by the cBM positions at time $t>0$ with conditional distribution $$\label{defn:Bernoulli}
{\mathcal{L}}\left(\big(\chi_y\big)_{y\in{{\mathbf Y}}_t^{{\mathbf x}}}\,\big|\,{{\mathbf Y}}^{{\mathbf x}}\right)=\bigotimes_{y\in{{\mathbf Y}}_t^{{\mathbf x}}}\mathrm{Ber}\left(u(y)\right).$$ We suppress the dependence on $u$ in the notation for these random variables.[^4] Recalling our notation , note that for $i\ne j$ the random variables $\chi_{Y_t^{{\scriptscriptstyle{({x_i}})}}}$ and $\chi_{Y_t^{{\scriptscriptstyle{({x_j}})}}}$ are either identical or independent, depending on whether or not the Brownian motions starting from $x_i$ and $x_j$ have coalesced up to time $t$.
\[thm:characterization\_entrance\_laws\] Let ${\mathcal{D}}$ be endowed with the topology introduced in Section \[sec:topology\].
- The semigroup $(T_t)_{t\ge0}$ on ${\mathcal{V}}^d$ defined in can be extended to a Feller semigroup $(\hat T_t)_{t\ge0}$ on ${\mathcal{V}}$ such that $$\label{eq:concentration}
\text{for all }v\in{\mathcal{V}}\text{ and }t>0:\; \hat T_t(v;\cdot)\text{ is concentrated on }{\mathcal{V}}^d.$$ The corresponding Feller process, which we denote by $(V_t)_{t\ge0}$, has continuous paths [ and is characterized by the following ‘moment duality’: writing $V_t=[U_t]=\{U_t,1-U_t\}\in{\mathcal{V}}$, we have for each $v=[u]\in{\mathcal{V}}$, $n\in{\mathbb{N}}$ and Lebesgue-almost all ${{\mathbf x}}=(x_1,\ldots,x_{2n})\in{\mathbb{R}}^{2n}$ $$\begin{aligned}
{\begin{aligned}}\label{eq:duality_3}
{\mathbb{P}}_v\left(\bigcap_{i=1}^n\{U_t(x_{2i-1})=U_t(x_{2i})\}\right)={\mathbb{P}}\left(\bigcap_{i=1}^n\big\{\chi_{Y_t^{{\scriptscriptstyle{({x_{2i-1}}})}}}=\chi_{Y_t^{{\scriptscriptstyle{({x_{2i}}})}}}\big\}\right),\qquad t>0,
{\end{aligned}}\end{aligned}$$ where $({{\mathbf Y}}_t^{{{\mathbf x}}})_{t\ge0}$ is a system of cBMs starting from ${{\mathbf x}}$ and [the Bernoulli random variables $\chi_{Y_t^{{\scriptscriptstyle{({x_i}})}}}$ are]{} as in . ]{}
- There is a bijective correspondence between probability entrance laws $\mu=(\mu_t)_{t>0}$ for the semigroup $(P_t)_{t\ge0}$ of aBMs on ${\mathcal{D}}$ and probability measures $\nu$ on ${\mathcal{V}}$, given by the formula $$\label{eq:correspondence}
\mu_t=\nu \hat T_t\circ{\mathcal{I}}^{-1}={\mathcal{L}}\left({\mathcal{I}}(V_t)\,|\,{\mathbb{P}}_{\nu}\right),\qquad t>0.$$ [ For $\nu=\delta_v$ with $v=[u]\in{\mathcal{V}}$, the corresponding entrance law $\mu$ is characterized by $$\label{eq:correspondence_0}
{\mathbb{P}}_{\mu}\left(\bigcap_{i=1}^n\{|{{\mathbf X}}_t\cap[x_{2i-1},x_{2i}]|\text{ is even}\}\right)={\mathbb{P}}\left(\bigcap_{i=1}^n\big\{\chi_{Y_t^{{\scriptscriptstyle{({x_{2i-1}}})}}} = \chi_{Y_t^{{\scriptscriptstyle{({x_{2i}}})}}}\big\}\right)$$ for all $t>0$, $n\in{\mathbb{N}}$ and ${{\mathbf x}}=(x_1,\ldots,x_{2n})\in{\mathbb{R}}^{2n,\uparrow}$. ]{}
Note that since replacing $u$ by $1-u$ is equivalent (in distribution) to replacing $\chi_{Y_t^{{\scriptscriptstyle{({x_i}})}}}$ by $1-\chi_{Y_t^{{\scriptscriptstyle{({x_i}})}}}$, the RHS of and depends indeed only on the equivalence class $v=[u]\in{\mathcal{V}}$. Also note that if $u$ is $\{0,1\}$-valued, we can choose $\chi_{Y_t^{{\scriptscriptstyle{({x_i}})}}}\equiv u(Y_t^{{\scriptscriptstyle{({x_i}})}})$ and so in this case reads $$\label{eq:border}
{\mathbb{P}}_{\mu}\left(\bigcap_{i=1}^n\left\{|{{\mathbf X}}_t\cap[x_{2i-1},x_{2i}]|\text{ is even}\right\}\right)={\mathbb{P}}\left(\bigcap_{i=1}^n\{u(Y_t^{{\scriptscriptstyle{({x_{2i-1}}})}}) = u(Y_t^{{\scriptscriptstyle{({x_{2i}}})}})\}\right).$$ We observe that this formula can be interpreted as an analogue in continuous space of the so-called *border equation* characterizing annihilating random walks on ${\mathbb{Z}}$, see e.g. [@BG80 Sec. 2].
Our next result clarifies the question raised in the introduction concerning different approximations of the real line by asymptotically dense subsets. In particular, it shows that each entrance law for aBMs can be approximated by a sequence of (random) initial conditions in ${\mathcal{D}}$.
\[thm:convergence\] Let ${\mathcal{D}}$ be endowed with the topology introduced in Section \[sec:topology\], and let $(V_t)_{t\ge0}$ denote the Feller process from Theorem \[thm:characterization\_entrance\_laws\]. Let $(\mu^{{\scriptscriptstyle{({n}})}})_{n\in{\mathbb{N}}}$ be a sequence of probability measures on ${\mathcal{D}}$, and consider the corresponding sequence of aBM processes started according to the (random) initial condition $\mu^{{\scriptscriptstyle{({n}})}}$. Then ${\mathcal{L}}\big(({{\mathbf X}}_t)_{t>0}\,\big|\,{\mathbb{P}}_{\mu^{{\scriptscriptstyle{({n}})}}}\big)$ converges weakly in ${\mathcal{C}}_{(0,\infty)}({\mathcal{D}})$ iff the sequence $(\mu^{{\scriptscriptstyle{({n}})}}\circ{\mathcal{I}})_{n\in{\mathbb{N}}}$ of probability measures on ${\mathcal{V}}^d$ converges weakly to some probability measure $\nu$ on ${\mathcal{V}}$, in which case $$\label{eq:convergence}
\lim_{n\to\infty}{\mathcal{L}}\big(({{\mathbf X}}_t)_{t>0}\,\big|\,{\mathbb{P}}_{\mu^{{\scriptscriptstyle{({n}})}}}\big)={\mathcal{L}}\big(({\mathcal{I}}(V_t))_{t>0}\,\big|\, {\mathbb{P}}_{\nu}\big)\quad\text{on }{\mathcal{C}}_{(0,\infty)}({\mathcal{D}}).$$ Moreover, for any entrance law $(\mu_t)_{t>0}$ for the semigroup $(P_t)_{t\ge0}$ of aBMs there exists a sequence $(\mu^{{\scriptscriptstyle{({n}})}})_{n\in{\mathbb{N}}}$ of probability measures on ${\mathcal{D}}$ such that $$\mu_t=\lim_{n\to\infty}\mu^{{\scriptscriptstyle{({n}})}}P_t,\qquad t>0.$$
To illustrate Theorems \[thm:characterization\_entrance\_laws\] and \[thm:convergence\], we give various examples showing the effect of different ways of approximating increasingly dense initial conditions for aBMs, see also Figure \[fig:simulations\].
- First, consider ${{\mathbf x}}_n=\frac1n{\ensuremath{\mathbb{Z}}}$. Clearly ${\mathcal{I}}^{-1}({{\mathbf x}}_n)$ converges to $[\frac12]$ in ${\mathcal{V}}$, and hence by Theorem \[thm:convergence\] the system of aBMs starting from ${{\mathbf x}}_n$ converges. We have the same limit when ${{\mathbf x}}_n$ is any other regularly spaced lattice with mesh going to zero as $n\to\infty$, or when ${{\mathbf x}}_n$ is the realisation of a Poisson point process of intensity $n$, [as in Figure \[fig:simulations\](b)]{}. These approximations give the ‘maximal’ entrance law considered in [@TZ11].
- In the example ${{\mathbf x}}_n=\frac1n{\ensuremath{\mathbb{Z}}}+\{0,\frac1{n^2}\}$ we still have convergence of ${\mathcal{I}}^{-1}({{\mathbf x}}_n)$ in ${\mathcal{V}}$, but the limit is $[0]$, which is degenerate and corresponds to the empty system. So indeed in the limit the close-by pairs have annihilated and there are no surviving aBMs. More generally, we have the same limit when ${{\mathbf x}}_n=\frac1n{\ensuremath{\mathbb{Z}}}+\{0,\frac1{n^\alpha}\}$ for some $\alpha>1$.
- We can also consider ${{\mathbf x}}_n=\frac1n{\ensuremath{\mathbb{Z}}}+\{0,\frac1{4n}\}$, where ${\mathcal{I}}^{-1}({{\mathbf x}}_n)$ converges to $[\frac14]$ in ${\mathcal{V}}$, which is different from $[\frac12]$. This is an example of an entrance law where we still start aBMs everywhere on the real line just as in $[\frac12]$, but the system ‘comes down from infinity’ in a different way, giving rise to a different law of the aBMs.
[0.48]{}
graphics \[xmin=0,xmax=100,ymin=0,ymax=70\] [aBM\_grey\_600x6000\_12.png]{};
[0.48]{}
graphics \[xmin=0,xmax=100,ymin=0,ymax=70\] [aBM\_grey\_600x6000\_ppp.png]{};
[0.48]{}
graphics \[xmin=0,xmax=100,ymin=0,ymax=70\] [aBM\_grey\_600x6000\_110.png]{};
[0.48]{}
graphics \[xmin=0,xmax=100,ymin=0,ymax=70\] [aBM\_grey\_600x6000\_14.png]{};
- As a final example we look at a sequence ${{\mathbf x}}_n\in{\mathcal{D}}$ such that ${\mathcal{I}}^{-1}({{\mathbf x}}_n)$ does not converge in ${\mathcal{V}}$: Let $x_n\in{\mathbb{R}}$ be a sequence converging monotone from below to some $a\in {\ensuremath{\mathbb{R}}}$. We put ${{\mathbf x}}_n=\{x_1,\ldots,x_n\}$ and write ${\mathcal{I}}^{-1}({{\mathbf x}}_n)=[ u_n]$ with $u_n\in{\mathcal{M}}_1^d({\mathbb{R}})$. Going from ${{\mathbf x}}_n$ to ${{\mathbf x}}_{n+1}$ adds the single point $x_{n+1}$, and we can choose the support of $u_{n+1}$ to remain fixed to the left of $x_{n+1}$, but such that it changes to the right of $x_{n+1}$. Then for any test function $\phi$ which is supported both to the left and to the right of $a$, the sequence $\langle u_n,\phi\rangle$ does not converge. However, if we add points in pairs, then the support of the induced measure remains unchanged except for the interval between the two added points, whose length goes to 0. Hence ${\mathcal{I}}^{-1}({{\mathbf x}}_{2n})$ and ${\mathcal{I}}^{-1}({{\mathbf x}}_{2n+1})$ converge to two distinct limit points. This is not surprising, since aBMs are parity preserving, and if we start with an even number eventually all will annihilate, while if we start with an odd number there will be a single surviving Brownian motion. However, if we extend the example to two sequences $x_n \uparrow a$ and $y_n \downarrow b$, $a<b$ and ${{\mathbf x}}_n=\{x_1,y_1,\ldots,x_n,y_n\}$, then the number of starting points is always even, but still ${\mathcal{I}}^{-1}({{\mathbf x}}_n)$ does not converge in ${\mathcal{V}}$. Note that we needed here that the sequence $x_n$ converges to a finite point $a\in {\ensuremath{\mathbb{R}}}$. If $a=\infty$, the above argument does not work and in fact ${\mathcal{I}}^{-1}({{\mathbf x}}_n)$ does converge in ${\mathcal{V}}$.
Results on $n$-point densities {#sec:densities}
------------------------------
In this subsection, we turn to the $n$-particle density function for aBMs, which is defined as follows: If $\mu=(\mu_t)_{t>0}$ is an entrance law for the semigroup $(P_t)_{t\ge0}$, the corresponding $n$-point density is given by $$\begin{aligned}
\label{eq:density}
p_{\mu}(t,{{\mathbf x}})&:=
\lim_{\epsilon\to0}\frac{1}{(2\epsilon)^n}\,{\ensuremath{\mathbb{P}}}_{\mu}\left(\bigcap_{i=1}^n\{{{\mathbf X}}_t\cap[x_i-\epsilon,x_i+\epsilon]\ne\emptyset\}\right),\end{aligned}$$ for ${{\mathbf x}}=(x_1,\ldots,x_n)\in {\ensuremath{\mathbb{R}}}^{n,\uparrow},\;t>0$. See e.g. [@MRTZ2006 Appendix B] for the existence of this density.
For $n=1$, the $1$-point density can be computed explicitly:
\[thm:density\] Let $v=[u]\in{\mathcal{V}}$ and consider the entrance law corresponding to $\nu:=\delta_{[u]}$ in view of Thm. \[thm:characterization\_entrance\_laws\]. Then the $1$-particle density function is given by $$\begin{aligned}
\label{eq:1-point-density}
p_{[u]}(t,x) = \frac{1}{2\pi t^{2}}\, \int_{{\ensuremath{\mathbb{R}}}^2}u(x+y_1)(1-u(x+y_2))|y_2-y_1|e^{-\frac{|{{\mathbf y}}|^2}{2t}}d{{\mathbf y}},\qquad x\in {\ensuremath{\mathbb{R}}},\;t>0.\end{aligned}$$
Observe that for the class of *homogeneous* entrance laws parametrized by $[\lambda]\in{\mathcal{V}}$, $\lambda\in[0,\frac12]$, the expression for the one-point density simplifies to $$p_{[\lambda]}(t,x) = \frac{2\lambda(1-\lambda)}{\sqrt{\pi t}}.$$
In particular, for the ‘maximal’ entrance law for aBMs considered in [@TZ11], corresponding to $\lambda=\frac{1}{2}$, we have $$p_{[\frac12]}(t,x) = \frac{1}{2\sqrt{\pi t}},$$ which indeed clearly maximizes the one-point density among all homogeneous entrance laws. Note that (as is to be expected from the thinning relation) this is half the density under the maximal entrance law for cBMs, compare [@SSS17 Prop. 2.7].
However, non-homogeneous entrance laws can achieve bigger densities. For example, [ if we choose $u:={\ensuremath{\mathbbm{1}}}_{{\mathbb{R}}^-}$, then the entrance law $\delta_{[{\ensuremath{\mathbbm{1}}}_{{\mathbb{R}}^-}]}$ ]{} corresponds to a single Brownian motion starting at the origin, for which we have $$p_{[{\ensuremath{\mathbbm{1}}}_{{\mathbb{R}}^-}]}(t,x)=\frac{1}{\sqrt{2\pi t}}e^{-\frac{x^2}{2t}},$$ and in particular $$p_{[{\ensuremath{\mathbbm{1}}}_{{\mathbb{R}}^-}]}(t,0)= \frac{1}{\sqrt{2\pi t}}>\frac{1}{2\sqrt{\pi t}}=p_{[\frac12]}(t,0).$$ This phenomenon is not limited to entrance laws which do not start densely: For $\epsilon\in(0,\frac12)$, consider the entrance law [ corresponding to $\delta_{[u]}$ with $$u:=\epsilon+(1-2\epsilon){\ensuremath{\mathbbm{1}}}_{{\mathbb{R}}^-}.$$ ]{} Here ${\mathcal{I}}([u])={\ensuremath{\mathbb{R}}}$, but $u(x)\to{\ensuremath{\mathbbm{1}}}_{{\mathbb{R}}^-}$ uniformly as $\epsilon\to0$, and by $p_{\bullet}(t,x)$ is continuous in the uniform topology, so that $p_{[u]}(t,0)>p_{[\frac12]}(t,0)$ for $\epsilon$ small enough. We conclude that a more appropriate name for the entrance law corresponding to $[\frac12]\in{\mathcal{V}}$ and discussed in [@TZ11] would be ‘maximal homogeneous’.
Turning to the case $n\ge2$, we have the following result:
\[prop:density-n\] Let $v=[u]\in{\mathcal{V}}$ and consider the entrance law corresponding to $\nu:=\delta_{[u]}$ in view of Thm. \[thm:characterization\_entrance\_laws\]. Then for each $n\in{\mathbb{N}}$, ${{\mathbf x}}=(x_1,\ldots,x_n)\in{\mathbb{R}}^{n,\uparrow}$ and $t>0$, we have $$\begin{aligned}
\label{eq:n-point-density}
p_{[u]}(t,{{\mathbf x}})
&= \lim_{{\varepsilon}\downarrow0} \frac{1}{(2{\varepsilon})^{n}}\,{\mathbb{P}}\bigg(\bigcap_{i=1}^n \{\chi_{Y^{{\scriptscriptstyle{({x_i-\epsilon}})}}_t}\ne \chi_{Y^{{\scriptscriptstyle{({x_i+\epsilon}})}}_t} \}\bigg),\end{aligned}$$ where $({{\mathbf Y}}_t^{(x_1-\epsilon,\,x_1+\epsilon,\ldots,x_n-\epsilon,\,x_n+\epsilon)})_{t\ge0}$ is a system of cBMs and the Bernoulli random variables $\chi_{Y_t^{{\scriptscriptstyle{({x_i\pm\epsilon}})}}}$ are as in .
\[rem:not\_explicit\]
- For $n=1$, the event on the RHS of simplifies to $$\{\tau^{{\scriptscriptstyle{({x,\epsilon}})}}>t\}\cap\{\chi_{Y_t^{{\scriptscriptstyle{({x-\epsilon}})}}}\ne\chi_{Y_t^{{\scriptscriptstyle{({x+\epsilon}})}}}\},$$ where $\tau^{{\scriptscriptstyle{({x,\epsilon}})}}$ is the coalescence time of the two Brownian motions. The $1$-point density can then be obtained by using the distribution of two Brownian motions conditioned not to collide up to time $t$, see the proof of Thm. \[thm:density\].
- Note that for a homogeneous entrance law $v=[\lambda]$ with $\lambda\in(0,1)$ constant, the RHS of can be simplified: For $\epsilon$ small enough, successive terms in the intersection are either independent or share a Bernoulli random variable via coalesced Brownian motions. Partitioning the index set $\{1,\ldots,n\}$ into $K$ blocks $\{i_k,\ldots,i_{k+1}-1\}$ (with $i_1:=1$ and $i_{K+1}:=n+1$) based on the coalescence structure, so that different blocks are independent, simplifies to $$\begin{aligned}
\label{eq:n-point-density-2}{\begin{aligned}}p_{[\lambda]}(t,{{\mathbf x}}) &= \lim_{{\varepsilon}\downarrow0} \frac{1}{(2{\varepsilon})^{n}}\,{\mathbb{E}}\left[{\ensuremath{\mathbbm{1}}}_{D_t^{{\scriptscriptstyle{({\epsilon}})}}}\,\prod_{k=1}^K (\lambda(1-\lambda))^{\lfloor\tfrac{1+i_{k+1}-i_{k}}2\rfloor}(1+{\ensuremath{\mathbbm{1}}}_{\{i_{k+1}-i_{k} \text{ is odd}\}})\right]\\
&=
\lim_{{\varepsilon}\downarrow0} \frac{1}{(2{\varepsilon})^{n}}\,{\mathbb{E}}\left[{\ensuremath{\mathbbm{1}}}_{D_t^{{\scriptscriptstyle{({\epsilon}})}}}\, (\lambda(1-\lambda))^{\tfrac{K_{odd}+n}2}2^{K_{odd}}\right],
{\end{aligned}}\end{aligned}$$ where $D_t^{{\scriptscriptstyle{({\epsilon}})}}:=\bigcap_{i=1}^n \{ Y_t^{{\scriptscriptstyle{({x_i-\epsilon}})}} \neq Y_t^{{\scriptscriptstyle{({x_i+\epsilon}})}} \}$ and $K_{odd}:=\sum_{k=1}^K {\ensuremath{\mathbbm{1}}}_{\{i_{k+1}-i_{k} \text{ is odd}\}}$. In particular, for the ‘maximal’ entrance law $\lambda=\frac{1}{2}$ we have $$\begin{aligned}
p_{[\frac12]}(t,{{\mathbf x}})=&\lim_{{\varepsilon}\downarrow0} \frac{1}{(4{\varepsilon})^{n}}\,{\ensuremath{\mathbb{P}}}\left(D_t^{{\scriptscriptstyle{({\epsilon}})}}\right).\end{aligned}$$ We see again that becomes maximal for $\lambda=\frac{1}{2}$, thus the $n$-point density function is maximized by $\lambda=\frac{1}{2}$ in the class of homogeneous entrance laws, for any $n\in{\mathbb{N}}$.
The expression for the $n$-point density function is deceivingly short. In fact, as we have just seen in Remark \[rem:not\_explicit\], it involves a lot of combinatorial effort to disentangle the effect of the various ways the Brownian motions can have coalesced.
However, we can give a more tractable representation of a ‘thinned’ version of the $n$-point density as follows: Fix ${{\mathbf x}}\in{\mathcal{D}}$. The discrete closed set ${{\mathbf x}}$ can be partitioned into two disjoint subsets ${{\mathbf x}}={{\mathbf x}}^{{\scriptscriptstyle{({1}})}}\cup{{\mathbf x}}^{{\scriptscriptstyle{({2}})}}$ so that points in ${{\mathbf x}}$ are alternating between ${{\mathbf x}}^{{\scriptscriptstyle{({1}})}}$ and ${{\mathbf x}}^{{\scriptscriptstyle{({2}})}}$ and such that either $\sup({{\mathbf x}}) \in {{\mathbf x}}^{{\scriptscriptstyle{({1}})}}$ or otherwise $\inf({{\mathbf x}}\cap [0,\infty)) = \inf({{\mathbf x}}^{{\scriptscriptstyle{({1}})}} \cap[0,\infty))$. Denote by ${{\mathbf x}}^{thin}$ the random subset of ${{\mathbf x}}$ equalling either ${{\mathbf x}}^{{\scriptscriptstyle{({1}})}}$ or ${{\mathbf x}}^{{\scriptscriptstyle{({2}})}}$ with probability $\frac12$.
Now for an entrance law $\mu=(\mu_t)_{t>0}$, we define the thinned $n$-point density as $$\begin{aligned}
p_\mu^{thin}(t,{{\mathbf x}}):= \lim_{\epsilon\to0}\frac{1}{(2\epsilon)^n}\,{\ensuremath{\mathbb{P}}}_{\mu}\left(\bigcap_{i=1}^n\{{{\mathbf X}}^{thin}_t\cap[x_i-\epsilon,x_i+\epsilon]\ne\emptyset\}\right),\qquad {{\mathbf x}}\in {\ensuremath{\mathbb{R}}}^{n,\uparrow},\;t>0,\end{aligned}$$ where ${{\mathbf X}}_t^{thin}$ is the random subset of ${{\mathbf X}}_t$ obtained via thinning as defined above.
\[thm:n-point-density-thinned\] Let $v=[u]\in{\mathcal{V}}$ and consider the entrance law corresponding to $\nu:=\delta_{[u]}$ in view of Thm. \[thm:characterization\_entrance\_laws\]. Then for ${{\mathbf x}}\in{\ensuremath{\mathbb{R}}}^{n,\uparrow}$ and $t>0$, we have $$\begin{aligned}
\label{eq:n-density}
&p_{[u]}^{thin}(t,{{\mathbf x}}) \\
&= \frac{q(t,{{\mathbf x}})}{2} \,\lim_{\epsilon\downarrow0}{\mathbb{E}}\bigg[\prod_{k=1}^n u( B^{{\scriptscriptstyle{({x_k-\epsilon}})}}_t)(1-u(B_t^{{\scriptscriptstyle{({x_k+\epsilon}})}})) + \prod_{k=1}^n (1-u(B_t^{{\scriptscriptstyle{({x_k-\epsilon}})}}))u(B^{{\scriptscriptstyle{({x_k+\epsilon}})}}_t) \,\bigg|\, \tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t\bigg],\end{aligned}$$ where $B$ is a Brownian motion in ${\mathbb{R}}^{2n}$ starting from and indexed by $(x_1-\epsilon,x_1+\epsilon,\ldots,x_n-\epsilon,x_n+\epsilon)$, $$\tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}:=\inf\{t>0\,|\,B_t^{{\scriptscriptstyle{({y}})}}=B_t^{{\scriptscriptstyle{({z}})}}\text{ for some }y\neq z\}$$ is the first collision time of any pair of coordinates, and $$\begin{aligned}
\label{eq:q_t}
q(t,{{\mathbf x}}):=\lim_{\epsilon\to0}\frac{1}{(2\epsilon)^n}\,{\ensuremath{\mathbb{P}}}_{}(\tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t).\end{aligned}$$
We recall that [@TZ11] show that the $n$-point densities for aBMs started in the ‘maximal homogeneous’ entrance law are given in terms of Pfaffians. It would be interesting to make the connection to our formulae, which however does not seem to be completely straight-forward, see also Remark \[rem:not\_explicit\].
The continuous-space voter model {#sec:cSSM}
================================
The proof of Theorem \[thm:characterization\_entrance\_laws\] (the characterization of entrance laws for annihilating Brownian motions) in Section \[sec:proofs\] below relies on a close connection of aBMs to what we call the *continuous-space voter model*. This section is devoted to a survey explaining this connection, which is also of independent interest. We will not give proofs but refer to the existing literature, commenting on necessary modifications when appropriate.
We start by recalling the classical (nearest-neighbor) voter model on ${\ensuremath{\mathbb{Z}}}^d$: This is a Markov process $(\eta_t)_{t\ge0}$ taking values in $\{0,1\}^{{\ensuremath{\mathbb{Z}}}^d}$ such that $$\label{eq:voter-discrete}
\eta(x)\text{ flips to }1-\eta(x) \text{ at rate }\frac{1}{2d}\sum_{y:|y-x|=1}{\ensuremath{\mathbbm{1}}}_{\{\eta(y)\ne \eta(x)\}},\qquad x\in{\mathbb{Z}}^d.$$ As is well known, it is characterized by the following *moment duality:* For all $\eta\in\{0,1\}^{{\ensuremath{\mathbb{Z}}}^d}$ and finite subsets $A\subset{\ensuremath{\mathbb{Z}}}^d$, we have $$\label{eq:moment-duality-discrete}
{\ensuremath{\mathbb{E}}}_{\eta}\Bigg[\prod_{x\in A}\eta_t(x)\Bigg]={\ensuremath{\mathbb{E}}}_{A}\Bigg[\prod_{x\in A_t} \eta(x)\Bigg],\qquad t\ge0,$$ where $(A_t)_{t\ge0}$ denotes a (set-valued) system of (instantaneously) coalescing nearest-neighbor random walks starting from $A$. See [@Liggett85] for this duality and for further background on the discrete voter model.
In view of this, a continuous-space analogue of the voter model should be a Markov process $(u_t)_{t\ge0}$ with $u_t(x)\in\{0,1\}$ for all $x\in{\mathbb{R}}$ which is characterized by a similar moment duality, but where the coalescing random walks are replaced by coalescing Brownian motions. Indeed, such a model was first introduced by [@Evans1997] in a much more general context and further discussed in [@Donnellyetal2000] and [@Zhou2003], where it is however called a *continuum-sites stepping-stone model*. The following result is essentially contained as a special case (only two types, Brownian migration on ${\mathbb{R}}$) in [@Evans1997 Thm. 4.1, Prop. 5.1] and [@Donnellyetal2000 Cor. 7.3]:
\[thm:voter\] There exists a unique Feller semigroup $(Q_t)_{t\ge0}$ on ${\mathcal{M}}_1({\mathbb{R}})$ such that the corresponding Feller process $(u_t)_{t\ge0}$ is characterized by the following moment duality: For all $u\in{\mathcal{M}}_1({\mathbb{R}})$ and $n\in{\mathbb{N}}$, we have for Lebesgue-almost all ${{\mathbf x}}=(x_1,\ldots,x_n)\in{\mathbb{R}}^n$ $$\label{eq:duality_pointwise}
{\mathbb{E}}_{u}\bigg[\prod_{i=1}^nu_t(x_i)\bigg]
={\mathbb{E}}\bigg[\prod_{y\in{{\mathbf Y}}^{{\mathbf x}}_t}u(y)\bigg],
\qquad t\ge0,$$ where ${{\mathbf Y}}^{{\mathbf x}}$ denotes a system of coalescing Brownian motions starting from ${{\mathbf x}}$. For each initial condition $u\in{\mathcal{M}}_1({\mathbb{R}})$, the process $(u_t)_{t\ge0}$ has continuous sample paths, and for each fixed $t>0$ we have, almost surely under ${\mathbb{P}}_u$, $$\label{eq:sep_types}
u_t(x)\in\{0,1\}\qquad\text{for Lebesgue-almost all }
x\in{\ensuremath{\mathbb{R}}}.$$
In view of the ‘separation of types’-property as well as the analogous form of the moment dualities and , we prefer to call the Feller process $(u_t)_{t\ge0}$ from Theorem \[thm:voter\] the *continuous-space voter model*, and will denote it by $\mathrm{CSVM}$ in the following. If we refer to a particular initial condition $u\in{\mathcal{M}}_1({\mathbb{R}})$, we write $\mathrm{CSVM}_u$. Note that implies that the model is symmetric under exchange of $u$ and $1-u$, in the sense that $$\label{eq:symmetric}
{\mathcal{L}}\left((1-u_t)_{t\ge0})\,|\,{\mathbb{P}}_{u}\right)={\mathcal{L}}\left((u_t)_{t\ge0})\,|\,{\mathbb{P}}_{1-u}\right),\qquad u\in{\mathcal{M}}_1({\mathbb{R}}).$$ The qualification ‘Lebesgue-almost all’ in Theorem \[thm:voter\] is necessary since the process $(u_t)_{t\ge0}$ is *measure-valued* (recall that on the state space ${\mathcal{M}}_1({\mathbb{R}})$, we use the topology of vague convergence). However, we will see below that a version of the densities $u_t$ can be chosen such that the moment duality holds for all ${{\mathbf x}}\in{\mathbb{R}}^n$, even in a pathwise sense and not only in expectation (see ). In particular, $u_t(x)$ is a Bernoulli random variable with parameter ${\mathbb{E}}_x[u(B_t)]$ for each fixed $t>0$ and $x\in{\mathbb{R}}$, where $B$ is a standard Brownian motion. Moreover, we will see that can be strengthened to a much stronger clustering property, see Thm. \[thm:interfaces\].
There are several possible constructions for CSVM. In [@Evans1997], Evans constructed the model directly from (a weak form of) the moment duality , even for much more general particle motions than Brownian motion on ${\mathbb{R}}$, and for an *uncountable* type space instead of the two-type case we consider here. In several later papers, CSVM was shown to arise as the limit of various other discrete- or continuous-space models. It was first observed in [@AS11 p. 794] (although without a formal proof) that the discrete one-dimensional voter model converges to $\mathrm{CSVM}$ under diffusive space/time-rescaling. Later, CSVM was obtained as the scaling limit of rescaled spatial $\Lambda$-Fleming-Viot processes on ${\mathbb{R}}$ (see [@BEV13 Thm. 1.1]), of rescaled interacting Fleming-Viot processes on ${\mathbb{Z}}$ under diffusive space/time-rescaling (see [@GSW16 Thms. 1.31, 1.32]), and of continuous-space *stepping stone models* $$\begin{aligned}
\label{eq:SSM}
\tfrac{\partial}{\partial t} u^{{\scriptscriptstyle{({\gamma}})}}_t(x) = \tfrac{1}{2}\Delta u^{{\scriptscriptstyle{({\gamma}})}}_t(x) + \sqrt{\gamma u^{{\scriptscriptstyle{({\gamma}})}}_t(x)(1-u^{{\scriptscriptstyle{({\gamma}})}}_t(x)) }\, \dot{W}_t(x),\qquad x\in{\mathbb{R}}\end{aligned}$$ as $\gamma\to\infty$ (see [@HOV15 Thm. 2.8a)]). The latter result makes sense in view of the fact that as first observed in [@Shiga86], the stepping stone model satisfies an analogous moment duality as in , but where the dual is a system of *delayed* cBMs, where two Brownian motions coalesce when their intersection local time exceeds an independent exponential random variable with parameter $\gamma$, and which clearly converges to a system of instantaneously coalescing Brownian motions as $\gamma\to\infty$.
The dynamics of the discrete voter model is specified explicitly by its flip rates . In contrast, in [@Evans1997] the CSVM-process $(u_t)_{t\ge0}$ was specified only indirectly via the moment duality . In analogy with the graphical representation of the discrete model (see e.g. [@Liggett85 Sec. III.6]), it is possible to give a more explicit *graphical* or *genealogical* construction of CSVM, as has been first observed in [@AS11] and then carried out rigorously in [@GSW16]. This however requires use of the (dual) *Brownian web*, a highly non-trivial object. We will briefly explain this graphical construction at the end of the present section.
An alternative ‘explicit’ construction of CSVM, which also provides the link to annihilating Brownian motions, is possible in terms of *interfaces*. Indeed, for the one-dimensional discrete model, it is quite easy to see (e.g. by the graphical representation) and was first observed in [@S78] that the dynamics of the discrete interface $$I_t:=\{x\in{\mathbb{Z}}\,|\, \eta_t(x)\ne \eta_t(x+1)\}$$ follows an *annihilating random walk*, and this provides an equivalent description of the voter model on ${\mathbb{Z}}$. It was shown in [@HOV15] that the analogous assertion holds in continuous space also. In other words, the interface model of CSVM is given by aBMs. But in continuous space, this is much more subtle since there we may start from initial conditions whose interface does not consist of well-separated points. In fact, even for an arbitrary initial condition $u\in{\mathcal{M}}_1({\mathbb{R}})$, the process ‘locally comes down from infinity’ immediately in the sense that almost surely, the set ${\mathcal{I}}(u_t)$ is discrete for each $t>0$, and the movement of the interface points for positive times is described in law by a system of aBMs.
A mathematically precise formulation is as follows: For $u\in{\mathcal{M}}_1^d({\mathbb{R}})$, consider a system of aBMs $({{\mathbf X}}_t)_{t\ge0}$ started from the discrete closed set ${\mathcal{I}}(u)\in{\mathcal{D}}$. The system $({{\mathbf X}}_t)_{t\ge0}$ induces a (random) partition of $[0,\infty)\times{\ensuremath{\mathbb{R}}}$, whose components are bounded by the closure $${\mathcal{J}}:= {\rm cl} \{ (t,{{\mathbf X}}_t)\, | \, t \in [0,\infty) \}$$ of the graphs of the annihilating paths. The path properties of the annihilating system ensure that the components of this partition can be assigned the value $0$ or $1$ in an alternating fashion, i.e. so that neighboring components have always different values. That is, we define a random mapping $$\label{eq:hat m}\hat m:{\mathcal{J}}^c\to\{0,1\}$$ with the property that it is locally constant and alternating on (each component of) ${\mathcal{J}}^c$. See Figure \[fig:colouring-1\] for an illustration.
![ An illustration of the construction of the process $(\hat U_t)_{t\ge0}\in{\mathcal{C}}_{[0,\infty)}({\mathcal{M}}^d_1({\mathbb{R}}))$ from an initial configuration $u$ with five interfaces and a system of aBMs starting from ${\mathcal{I}}(u)$. The density $\hat U_t(x)$ is equal to $1$ for all $(t,x)$ in the shaded area and is zero otherwise.[]{data-label="fig:colouring-1"}](colouring-1.pdf)
Of course, given the aBM path ${{\mathbf X}}$ there are exactly two possibilities to choose the mapping $\hat m$ on ${\mathcal{J}}^c$. However, it is determined by the choice of $\hat m(0,\cdot)$ at time zero. This enables us to define an ${\mathcal{M}}_1^d({\mathbb{R}})$-valued process $(\hat U_t)_{t\ge0}$ as follows: Given $u\in {\mathcal{M}}_1^d({\mathbb{R}})$, let $$\hat m(0,x):={\ensuremath{\mathbbm{1}}}_{{{\rm supp}}(u)}(x),\qquad x\in{\mathbb{R}},$$ then extend $\hat m(0, \cdot)$ to $[0,\infty)\times{\mathbb{R}}$ by the requirement that it is constant on the components of ${\mathcal{J}}^c$ as described above, and set $$\label{eq:defn_U_hat}\hat U_t(x):={\ensuremath{\mathbbm{1}}}_{\hat m(t,x)=1},\qquad x\in{\mathbb{R}},\; t\geq 0;$$ again see Figure \[fig:colouring-1\]. By the definition of the topology on ${\mathcal{M}}_1^d({\mathbb{R}})$, it is clear that the process $(\hat U_t)_{t\ge0}$ is a random element of ${\mathcal{C}}_{[0,\infty)}({\mathcal{M}}_1^d({\ensuremath{\mathbb{R}}}))$.
Then the following theorem is contained as a special case in [@HOV15 Thms. 2.12, 2.14]:
\[thm:interfaces\] Let $(u_t)_{t\ge0}$ be the $\mathrm{CSVM}_u$-process from Theorem \[thm:voter\].
- Suppose $u\in{\mathcal{M}}_1^d({\mathbb{R}})$, i.e. ${\mathcal{I}}(u)\in {\mathcal{D}}$. If we let the process $(\hat U_t)_{t\ge0}$ be defined as in above, then $$(u_t)_{t\ge0}\overset{d}=(\hat U_t)_{t\ge0}\quad\text{on }{\mathcal{C}}_{[0,\infty)}({\mathcal{M}}_1({\ensuremath{\mathbb{R}}})).$$
- Let $u\in{\mathcal{M}}_1({\mathbb{R}})$. Then, almost surely, we have ${\mathcal{I}}(u_{t})\in {\mathcal{D}}$ for all $t>0$. Moreover, for any $t_0>0$, the evolution of $(u_t)_{t\ge t_0}$ is given (in law) as in a) when started in $u_{t_0}$.
We note that [@Donnellyetal2000 Thm. 10.2] contains a somewhat analogous result for the corresponding continuous-space voter model with Brownian migration on the *torus*. (In that case, by compactness, the system comes down to *finitely* many interfaces immediately.) We also remark that [@Zhou2003] studies clustering behavior for the model with Brownian migration on the real line as in our case, but with infinitely many types as in [@Evans1997]. In particular, [@Zhou2003 Thm. 3.7] (when restricted to the two-types case) shows essentially that under homogeneous initial conditions $u_0\equiv u\in(0,1)$, for each *fixed* $t>0$, the interface ${\mathcal{I}}(u_t)$ is discrete almost surely. This is not strong enough to give rise to an ‘interface process’ as in Thm. \[thm:interfaces\].
1. Theorem \[thm:interfaces\] provides the crucial link between annihilating Brownian motions and the continuous-space voter model. In particular, we will see in the proof that for the process $(V_t)_{t \geq 0}$ from Theorem \[thm:characterization\_entrance\_laws\], if $V_0 = v = [u]$, then we have $(V_t)_{t\ge0} \stackrel{d}{=} ([u_t])_{t\ge0}$.
2. Write $(Q_t)_{t\ge0}$ for the semigroup of $(u_t)_{t\ge0}$ as in Thm. \[thm:voter\]. Then Theorem \[thm:interfaces\]b) implies in particular that $$\label{eq:clustering}
\text{for all }u\in{\mathcal{M}}_1({\mathbb{R}})\text{ and }t>0:\; Q_t(u;\cdot)\text{ is concentrated on }{\mathcal{M}}_1^d({\mathbb{R}}).$$
We conclude this section with an outline of the graphical construction of CSVM in terms of the (dual) Brownian web, first conceived in [@AS11] and later elaborated in [@GSW16]. For the analogous graphical representation of the discrete voter model, see e.g. [@Liggett85 Sec. III.6]. Our exposition will be non-technical; for background and technical details concerning the Brownian web, in particular the precise state space and topology involved, we refer to [@SSS17]. Actually, we will use the *double Brownian web* $({\mathcal{W}},\widehat{\mathcal{W}})$, which is a coupled construction of the Brownian web and its dual on the same probability space. That is, ${\mathcal{W}}=\{{\mathcal{W}}^{{\scriptscriptstyle{({t,x}})}}\}_{(t,x)\in{\mathbb{R}}^2}$ resp. $\widehat{\mathcal{W}}=\{\widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x}})}}\}_{(t,x)\in{\mathbb{R}}^2}$ is a (random) collection of paths indexed by the starting position $(t,x)$ and evolving forwards resp. backwards in time as coalescing Brownian motions, and such that almost surely, no path in ${\mathcal{W}}$ crosses any path in $\widehat{\mathcal{W}}$. More precisely, for each $(t,x)\in{\mathbb{R}}^2$, almost surely there is a unique path ${\mathcal{W}}^{{\scriptscriptstyle{({t,x}})}}$ starting from $(t,x)$ and not crossing any path in $\widehat{\mathcal{W}}$, and for any finite collection $(t_1,x_1),\ldots,(t_n,x_n)$ of starting points, $({\mathcal{W}}^{{\scriptscriptstyle{({t_1,x_1}})}},\ldots, {\mathcal{W}}^{{\scriptscriptstyle{({t_n,x_n}})}})$ is distributed as a system of $n$ coalescing Brownian motions, and analogously for $\widehat{\mathcal{W}}$.
Now the graphical construction of CSVM works as follows: Suppose first that $u\in{\mathcal{M}}_1^d({\mathbb{R}})$, so that we may assume $u(x)\in\{0,1\}$ for all $x\in{\mathbb{R}}$. Then we set $$u_t(x):=u(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}})\in\{0,1\},\qquad t\ge0,\; x\in{\mathbb{R}}.$$ This gives a coupled definition of the Bernoulli random variables $u_t(x)$ for all $(t,x)$ and has an interpretation in terms of *genealogies*: namely, the type of an ‘individual’ at time $t>0$ and site $x\in{\mathbb{R}}$ is determined by tracing back its genealogy along the backwards coalescing path $\widehat {\mathcal{W}}^{{\scriptscriptstyle{({t,x}})}}$ starting from $(t,x)$ down until time zero. (Note the analogy with the graphical representation of the discrete voter model.) For a general initial condition $u\in{\mathcal{M}}_1({\mathbb{R}})$, in order to define $u_t(x)$, one traces back the genealogy in the same way and then in the last step, conditionally on the realization of $\widehat{\mathcal{W}}$, one samples a type in $\{0,1\}$ according to a Bernoulli distribution with parameter $u(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}})$. Some care is needed to ensure that the resulting construction of $u_t(\cdot)$ is measurable in $x$.[^5] Observing that the set $$\widehat E:=\left\{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}}\,|\,t>0,\,x\in{\mathbb{R}}\right\}$$ is countable almost surely (see [@GSW16 p. 15]), we let $(U_n)_{n\in{\mathbb{N}}}$ be a sequence of i.i.d. uniform random variables on $[0,1]$, independent of $\widehat{\mathcal{W}}$, and given a realization of $\widehat{\mathcal{W}}$, we let $J:\widehat E\to{\mathbb{N}}$ be an enumeration of the elements of $\widehat E$. Then we put for $t>0$ and $x\in{\mathbb{R}}$ $$\label{eq:graphical-construction}
u_t(x):=\chi_{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}}}:=\begin{cases}1&\text{if }U_{J(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}})}\le u(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}}),\\0&\text{else}.\end{cases}$$
It is easy to check that conditionally on $\widehat{\mathcal{W}}$, the random variables $\chi_{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}}}$ (as a family indexed by the elements of $\widehat E$) are independent and Bernoulli distributed with parameter $u(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}})$, and that $u_t(\cdot)$ is measurable in $x$, hence yields a random element of ${\mathcal{M}}_1({\mathbb{R}})$.
In order to check that the process $(u_t)_{t\ge0}$ defined above agrees with the continuous-space voter model, note first that implies the moment duality in a pathwise sense, i.e. $$\label{eq:duality_pathwise}
\prod_{i=1}^n u_t(x_i)=\prod_{i=1}^n \chi_{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i}})}}}, \qquad t>0,\;{{\mathbf x}}=(x_1,\ldots,x_n)\in{\mathbb{R}}^n.$$ (In the terminology of [@JK14], we have a strong pathwise duality.) Taking expectations, we obtain immediately the moment duality , since $\{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_1}})}},\ldots,\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_n}})}}\}\overset{d}{=}{{\mathbf Y}}^{{\mathbf x}}_t$ for a system $({{\mathbf Y}}^{{\mathbf x}})_{t\ge0}$ of (forward) cBMs starting from ${{\mathbf x}}$. Because the moments determine the distribution, this shows that the one-dimensional marginals of $(u_t)_{t\ge0}$ defined in agree with those of CSVM. In order to conclude that the processes agree in distribution on path space, one needs to check that $(u_t)_{t\ge0}$ is a Markov process and has continuous paths; for a proof of this in a more general (infinitely-many-types) context, see [@GSW16 Thm. 1.27].
This pathwise ‘graphical’ construction of CSVM allows one to quickly see the assertions of Thm. \[thm:interfaces\]. For example, to see that the system locally comes down from infinity immediately, we fix $t>0$, $x\in{\mathbb{R}}$ and define $$J_t(x):=\{y\in{\mathbb{R}}\,|\, \widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,y}})}} = \widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}}\}.$$ The properties of $\widehat{\mathcal{W}}$ imply that $J_t(x)$ is a non-degenerate interval for each $x\in{\mathbb{R}}$, and by we have $u_t(y)=u_t(x)$ for all $y\in J_t(x)$. Thus considered as a measure, $u_t$ is supported on a countable collection of disjoint intervals. Moreover, the boundaries of these intervals cannot accumulate, since otherwise there would exist a finite interval containing at time $0$ infinitely many endpoints of backward coalescing paths starting at time $t>0$, contradicting the fact that these cBMs form a locally finite system at time $0$. In other words, we must have $u_t\in{\mathcal{M}}_1^d({\mathbb{R}})$, which is Thm. \[thm:interfaces\]b). Note that so far we used only the dual (‘backward’) Brownian web $\widehat{\mathcal{W}}$. However, by the coupling with the ‘forward’ Brownian web ${\mathcal{W}}$ it is quite easy to see the assertion of Thm. \[thm:interfaces\]a) that starting from $u\in{\mathcal{M}}_1^d({\mathbb{R}})$, interface points move as annihilating Brownian motions. For example, suppose that $u$ is supported on countably many disjoint compact intervals, i.e. $u=\sum_{k\in{\mathbb{Z}}}{\ensuremath{\mathbbm{1}}}_{[a_k,b_k]}$ with $a_k<b_k<a_{k+1}$ for all $k\in{\mathbb{Z}}$. Then by we have $u_t(x)
=\sum_{k=1}^n {\ensuremath{\mathbbm{1}}}_{J_t^{{\scriptscriptstyle{({k}})}}}(x)$, where $J_t^{{\scriptscriptstyle{({k}})}}:=\left\{x\in{\mathbb{R}}\,|\,\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}}\in[a_k,b_k]\right\}$. By the properties of $\widehat{\mathcal{W}}$, it is clear that each $J_t^{{\scriptscriptstyle{({k}})}}$ is a finite non-degenerate interval $J_t^{{\scriptscriptstyle{({k}})}}=[a_k(t),b_k(t)]$ provided it is not empty. We claim that for those indices $k$ such that $J_t^{{\scriptscriptstyle{({k}})}}\ne\emptyset$, we have $$a_k(t)={\mathcal{W}}_t^{{\scriptscriptstyle{({0,a_k}})}}\qquad\text{and}\qquad b_k(t)={\mathcal{W}}_t^{{\scriptscriptstyle{({0,b_k}})}}.$$ Indeed, this follows from the fact that no path in ${\mathcal{W}}$ can cross any path in $\widehat{\mathcal{W}}$. This shows that (locally) interface points move as Brownian motions as long as $J_t^{{\scriptscriptstyle{({k}})}}\ne\emptyset$. The coalescence time $\tau$ of the forward paths ${\mathcal{W}}^{{\scriptscriptstyle{({0,a_k}})}}$ and ${\mathcal{W}}^{{\scriptscriptstyle{({0,b_k}})}}$ is precisely the time from which on no backward path in $\widehat{\mathcal{W}}$ can end up in the interval $[a_k,b_k]$ at time zero, and so $J_t^{{\scriptscriptstyle{({k}})}}=\emptyset$ for $t\ge\tau$, which means that the interfaces corresponding to the $k$-th interval annihilate at time $\tau$. Obviously, analogous arguments work also for all other kinds of initial conditions in ${\mathcal{M}}_1^d({\mathbb{R}})$, for example when there are only finitely many interfaces and/or when the support of $u$ is unbounded in one or both directions.
We thus see that the graphical construction of CSVM sketched above allows for simple and elegant arguments in situations where the proofs would be more involved if one had to use only the moment duality . On the other hand, it relies on properties of the (dual) Brownian web, a highly non-trivial object. We will use the pathwise duality in some of the proofs in Section \[sec:proofs\] below, but emphasize that all results in this paper can also be proved without recourse to the Brownian web, using essentially only . Moreover, the moment duality technique is robust to some degree and can be generalized to situations where an analogous web construction does not (yet) exist. (Recall that [@Evans1997] allowed for much more general migration mechanisms than Brownian motion on ${\mathbb{R}}$, and [@HOV15] considered interface points which move as aBMs with a highly non-regular drift.)
Proofs of results {#sec:proofs}
==================
In this section, we prove our results stated above in Section \[sec:classification\].
For the proof of part a), recall that $(P_t)_{t\ge0}$ denotes the semigroup of annihilating Brownian motions on ${\mathcal{D}}$, and that the semigroup $(T_t)_{t\ge0}$ on ${\mathcal{V}}^d$ is defined as the image of $(P_t)_{t\ge0}$ under the inverse interface operator ${\mathcal{I}}^{-1}:{\mathcal{D}}\to{\mathcal{V}}^d$, see . The latter can be rewritten as $$\label{identity_semigroups}
T_t(v;f\circ{\mathcal{I}})=P_t({\mathcal{I}}(v);f),\qquad v\in{\mathcal{V}}^d,\;f\in{\mathcal{C}}_b({\mathcal{D}}).$$
On the other hand, recall that $(Q_t)_{t\ge0}$ denotes the Feller semigroup on ${\mathcal{M}}_1({\mathbb{R}})$ corresponding to the continuous-space voter process $(u_t)_{t\ge0}$ from Thm. \[thm:voter\]. Via the canonical quotient mapping $q:{\mathcal{M}}_1({\mathbb{R}})\to{\mathcal{V}}$, $(Q_t)_{t\ge0}$ factorizes to a Feller semigroup $$\label{qe:defn_Q}
\hat T_t(v;\cdot):=Q_t(u;\cdot)\circ q^{-1},\qquad v=[u]\in{\mathcal{V}},\;t\ge0$$ on the quotient space ${\mathcal{V}}$, and the corresponding Feller process $$(V_t)_{t\ge0}:=([u_t])_{t\ge0}$$ inherits the path continuity from $(u_t)_{t\ge0}$. Of course, for to make sense we need to check in particular that the definition does not depend on the choice of the representative $u$ or $1-u$ of the equivalence class $[u]$. But this (as well as the Feller property) follows easily from the symmetry . The property of the semigroup $(\hat T_t)_{t\ge0}$ follows immediately from the corresponding clustering property of $(Q_t)_{t\ge0}$.
In order to see that $(\hat T_t)_{t\ge0}$ as defined in is indeed an extension of $(T_t)_{t\ge0}$ as defined in , observe that for any $v\in{\mathcal{V}}^d$ we have by Theorem \[thm:interfaces\]a) that $$\label{representation_aBM}
{\mathcal{L}}\big(({{\mathbf X}}_t)_{t\ge0}\,\big|\,{\mathbb{P}}_{{\mathcal{I}}(v)}\big)={\mathcal{L}}\big(({\mathcal{I}}(V_t))_{t\ge0}\,\big|\, {\mathbb{P}}_{v}\big)\quad\text{on }{\mathcal{C}}_{[0,\infty)}({\mathcal{D}}).$$ But together with , this implies $$\begin{aligned}
{\begin{aligned}}T_t(v;f)&=P_t\big({\mathcal{I}}(v);f\circ{\mathcal{I}}^{-1}\big)={\mathbb{E}}_{{\mathcal{I}}(v)}\left[f\circ{\mathcal{I}}^{-1}({{\mathbf X}}_t)\right]={\mathbb{E}}_v[f(V_t)]
=\hat T_t(v;f)
{\end{aligned}}\end{aligned}$$ for each $v\in{\mathcal{V}}^d$, $f\in{\mathcal{C}}_b({\mathcal{V}}^d)$ and $t\ge0$.
The ‘moment duality’ formula follows directly from the graphical construction of the CSVM-process $(u_t)_{t\ge0}$, since $(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_{1}}})}},\ldots,\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_{2n}}})}})\overset{d}=(Y_t^{{\scriptscriptstyle{({x_1}})}},\ldots,Y_t^{{\scriptscriptstyle{({x_{2n}}})}})$ for a system $({{\mathbf Y}}^{{\mathbf x}})_{t\ge0}$ of (forward) cBMs starting from ${{\mathbf x}}$. It remains to show that this duality characterizes the law of $V_t$ on ${\mathcal{V}}$ for fixed $t>0$. To this end, we argue as follows: For $u\in{\mathcal{M}}_1({\mathbb{R}})$ and $x,y\in{\mathbb{R}}$, define $$\begin{aligned}
\label{eq:definition-h}
h_u(x,y)\equiv h_{1-u}(x,y):=u(x)(1-u(y))+(1-u(x))u(y). \end{aligned}$$ Note that $h_u(\cdot,\cdot)$ depends only on the equivalence class $[u]\in{\mathcal{V}}$, and that for $u\in{\mathcal{M}}_1^d({\mathbb{R}})$ we may assume $u(\cdot)\in\{0,1\}$ and thus $$\begin{aligned}
\label{eq:equality-h}
h_u(x,y)={\ensuremath{\mathbbm{1}}}_{\{u(x)\ne u(y)\}}\qquad\text{for } u\in{\mathcal{M}}_1^d({\mathbb{R}})\text{ and Lebesgue-almost all }x,y\in{\mathbb{R}}.\end{aligned}$$
Now given $n\in{\mathbb{N}}$ and $\phi_1,\ldots,\phi_{2n}\in{\mathcal{C}}_c({\mathbb{R}})$, we define $\Phi:=\phi_1\otimes\cdots\otimes\phi_{2n}\in{\mathcal{C}}_c({\mathbb{R}}^{2n})$ and a function $F_{\Phi}:{\mathcal{V}}\to{\mathbb{R}}$ by $$\begin{aligned}
F_{\Phi}(v)&:=\int_{{\mathbb{R}}^{2n}}\Phi({{\mathbf x}})\prod_{i=1}^n h_u(x_{2i-1},x_{2i})\,d{{\mathbf x}}\\
&=\prod_{i=1}^n\big(\langle u,\phi_{2i-1}\rangle\langle 1-u,\phi_{2i}\rangle+\langle u,\phi_{2i}\rangle\langle 1-u,\phi_{2i-1}\rangle\big),\qquad v=[u]\in{\mathcal{V}}. \end{aligned}$$ Note that $F$ is well-defined and continuous on ${\mathcal{V}}$, and the family of functions $${\mathcal{F}}:=\left\{F_{\Phi}(\cdot)\,|\,n\in{\mathbb{N}},\phi_i\in{\mathcal{C}}({\mathbb{R}})\text{ for } i=1,\ldots,2n\right\}\subseteq{\mathcal{C}}({\mathcal{V}})$$ is closed under multiplication and separates the points of ${\mathcal{V}}$. Therefore ${\mathcal{F}}$ is separating for probability laws on ${\mathcal{V}}$. Since $V_t\in{\mathcal{V}}^d$ for $t>0$ by , using we obtain $$\begin{aligned}
{\mathbb{E}}_v\left[F_{\Phi}(V_t)\right]
&=\int_{{\mathbb{R}}^{2n}}\Phi({{\mathbf x}})\,{\mathbb{P}}_v\left(\bigcap_{i=1}^n\big\{U_t(x_{2i-1})\ne U_t(x_{2i})\big\}\right) d{{\mathbf x}},\end{aligned}$$ showing that determines the law of $V_t$ on ${\mathcal{V}}$.
Thus part a) of Thm. \[thm:characterization\_entrance\_laws\] is proved.
For part b), let $\nu$ be any probability measure on ${\mathcal{V}}$ and define $\mu_t$ by . Then by we have for any $0<s<t$ and $f\in{\mathcal{C}}_b({\mathcal{D}})$ that $${\begin{aligned}}\mu_t(f)&=\nu\hat T_t(f\circ {\mathcal{I}})=\int_{{\mathcal{V}}^d}\hat T_{t-s}(\cdot;f\circ{\mathcal{I}})\, d(\nu\hat T_s)\\
&=\int_{{\mathcal{V}}^d}P_{t-s}({\mathcal{I}}(\cdot);f)\,d(\nu\hat T_s)=\int_{\mathcal{D}}P_{t-s}(\cdot,f)\,d\mu_s,{\end{aligned}}$$ showing that $\mu_t=\mu_sP_{t-s}$ and $(\mu_t)_{t>0}$ is an entrance law for the semigroup $(P_t)_{t\ge0}$ of aBMs. Conversely, suppose that $(\mu_t)_{t>0}$ is any such entrance law. Then a similar calculation shows that $\nu_t:=\mu_t\circ {\mathcal{I}}$ defines an entrance law for the semigroup $(\hat T_t)_{t\ge0}$ on ${\mathcal{V}}$. Since the latter is a Feller semigroup on a compact space, the entrance law $(\nu_t)_{t>0}$ is closable, i.e. there exists a probability measure $\nu$ on ${\mathcal{V}}$ such that $\nu_t=\nu\hat T_t$ for all $t>0$. In fact, let ${\mathcal{P}}({\mathcal{V}})$ denote the space of all probability measures on ${\mathcal{V}}$ endowed with the topology of weak convergence. Then ${\mathcal{P}}({\mathcal{V}})$ is itself compact, and thus there exists a sequence $t_n\downarrow0$ and $\nu\in{\mathcal{P}}({\mathcal{V}})$ such that $\nu_{t_n}\to\nu$ weakly as $n\to\infty$. Then for any $f\in{\mathcal{C}}_b({\mathcal{V}})$, $t>0$ and $n\in{\mathbb{N}}$ large enough we have by the Feller property that $$\nu_t(f)=\nu_{t_n}\hat T_{t-t_n}(f)\to \nu\hat T_t(f),$$ i.e. $\nu_t=\nu\hat T_t$, and we conclude that $\mu_t=\nu_t\circ{\mathcal{I}}^{-1}=\nu\hat T_t\circ{\mathcal{I}}^{-1}$. Moreover, the moment duality implies that different probability measures $\nu\ne\tilde\nu$ on ${\mathcal{V}}$ lead to different laws $$\label{eq:unique}{\mathcal{L}}\big(V_t\,|\,{\mathbb{P}}_{\nu}\big)\ne {\mathcal{L}}\big(V_t\,|\,{\mathbb{P}}_{\tilde\nu}\big)\qquad \text{for }t>0.$$ Thus the mapping defined by is indeed a bijection, and [ the claimed correspondence ]{} is established. [ For $\nu=\delta_v$ with $v=[u]\in{\mathcal{V}}$ and the corresponding entrance law $\mu$, implies $$\begin{aligned}
{\mathbb{P}}_{\mu}\left(\bigcap_{i=1}^n\{|{{\mathbf X}}_t\cap[x_{2i-1},x_{2i}]|\text{ is even}\}\right)&={\mathbb{P}}_{v}\left(\bigcap_{i=1}^n\{|{\mathcal{I}}(V_t)\cap[x_{2i-1},x_{2i}]|\text{ is even}\}\right)\end{aligned}$$ for each ${{\mathbf x}}\in{\mathbb{R}}^{2n,\uparrow}$ and $t>0$. Again writing $V_t=[U_t]=\{U_t,1-U_t\}$, we observe that on $\{{\mathcal{I}}(V_t)\cap{{\mathbf x}}=\emptyset\}$ (an event of full probability), we have that $|{\mathcal{I}}(V_t)\cap[x_{2i-1},x_{2i}]|$ is even iff $U_t(x_{2i-1})=U_t(x_{2i})$, for all $i=1,\ldots,n$. Thus by , the previous display equals $$\begin{aligned}
{\mathbb{P}}_v\left(\bigcap_{i=1}^n\{U_t(x_{2i-1}) = U_t(x_{2i})\}\right)={\mathbb{P}}\left(\bigcap_{i=1}^n\big\{\chi_{Y_t^{{\scriptscriptstyle{({x_{2i-1}}})}}} = \chi_{Y_t^{{\scriptscriptstyle{({x_{2i}}})}}}\big\}\right),\end{aligned}$$ establishing formula . ]{}
Suppose that $(\mu^{{\scriptscriptstyle{({n}})}})_{n\in{\mathbb{N}}}$ is a sequence of probability measures on ${\mathcal{D}}$ such that $\nu^{{\scriptscriptstyle{({n}})}}:=\mu^{{\scriptscriptstyle{({n}})}}\circ{\mathcal{I}}$ converges weakly to some probability measure $\nu$ on ${\mathcal{V}}$. Then by the Feller property of $(V_t)_{t\ge0}$, we have ${\mathcal{L}}\big((V_t)_{t\ge0}\,\big|\, {\mathbb{P}}_{\nu^{{\scriptscriptstyle{({n}})}}}\big)\to{\mathcal{L}}\big((V_t)_{t\ge0}\,\big|\, {\mathbb{P}}_{\nu}\big)$ weakly on ${\mathcal{C}}_{[0,\infty)}({\mathcal{V}})$, thus also $$\label{eq:proof_convergence}
{\mathcal{L}}\big((V_t)_{t>0}\,\big|\, {\mathbb{P}}_{\nu^{{\scriptscriptstyle{({n}})}}}\big)\to{\mathcal{L}}\big((V_t)_{t>0}\,\big|\, {\mathbb{P}}_{\nu}\big) \quad\text{weakly on }{\mathcal{C}}_{(0,\infty)}({\mathcal{V}}^d).$$ By the continuous mapping theorem and , it follows that $${\mathcal{L}}\big(({{\mathbf X}}_t)_{t>0}\,\big|\,{\mathbb{P}}_{\mu^{{\scriptscriptstyle{({n}})}}}\big)={\mathcal{L}}\big(({\mathcal{I}}(V_t))_{t>0}\,\big|\,{\mathbb{P}}_{\nu^{{\scriptscriptstyle{({n}})}}}\big)\to{\mathcal{L}}\big(({\mathcal{I}}(V_t))_{t>0}\,\big|\, {\mathbb{P}}_{\nu}\big)\quad\text{on }{\mathcal{C}}_{(0,\infty)}({\mathcal{D}}),$$ i.e. . Conversely, suppose that ${\mathcal{L}}\big(({{\mathbf X}}_t)_{t>0}\,\big|\,{\mathbb{P}}_{\mu^{{\scriptscriptstyle{({n}})}}}\big)={\mathcal{L}}\big(({\mathcal{I}}(V_t))_{t>0}\,\big|\, {\mathbb{P}}_{\mu^{{\scriptscriptstyle{({n}})}}\circ{\mathcal{I}}}\big)$ converges weakly in ${\mathcal{C}}_{(0,\infty)}({\mathcal{D}})$. Consider the sequence $\nu^{{\scriptscriptstyle{({n}})}}:=\mu^{{\scriptscriptstyle{({n}})}}\circ{\mathcal{I}}$ of probability measures on ${\mathcal{V}}^d$. Since ${\mathcal{V}}^d\subseteq{\mathcal{V}}$ and ${\mathcal{V}}$ is compact, this sequence is relatively compact w.r.t. the topology of weak convergence. Moreover, by continuous mapping also ${\mathcal{L}}\big((V_t)_{t>0}\,\big|\,{\mathbb{P}}_{\nu^{{\scriptscriptstyle{({n}})}}} \big)$ converges weakly in ${\mathcal{C}}_{(0,\infty)}({\mathcal{V}})$. But this implies (see ) that for any two limit points $\nu$ and $\tilde\nu$ of the sequence $(\nu^{{\scriptscriptstyle{({n}})}})_{n\in{\mathbb{N}}}$, we must have ${\mathcal{L}}\big((V_t)_{t>0}\,\big|\, {\mathbb{P}}_{\nu}\big)={\mathcal{L}}\big((V_t)_{t>0}\,\big|\, {\mathbb{P}}_{\tilde\nu}\big)$ on ${\mathcal{C}}_{(0,\infty)}({\mathcal{V}})$ and thus $\nu=\tilde\nu$ (recall ). Thus $(\nu^{{\scriptscriptstyle{({n}})}})_{n\in{\mathbb{N}}}$ must converge weakly to some probability measure $\nu$ on ${\mathcal{V}}$.
Finally, let $(\mu_t)_{t>0}$ be any probability entrance law for the semigroup $(P_t)_{t>0}$ of aBMs on ${\mathcal{D}}$. Let $\nu$ be the (unique) probability measure on ${\mathcal{V}}$ corresponding to the entrance law in view of Theorem \[thm:characterization\_entrance\_laws\]. Since ${\mathcal{V}}^d$ is dense in ${\mathcal{V}}$, there is a sequence $(\nu^{{\scriptscriptstyle{({n}})}})_{n\in{\mathbb{N}}}$ of probability measures concentrated on ${\mathcal{V}}^d$ such that $\nu^{{\scriptscriptstyle{({n}})}}\to\nu$ weakly as $n\to\infty$. Then putting $\mu^{{\scriptscriptstyle{({n}})}}:=\nu^{{\scriptscriptstyle{({n}})}}\circ {\mathcal{I}}^{-1}$ and again using the Feller property of $(\hat T_t)_{t\ge0}$ as well as , we get $$\mu_t=\nu\hat T_t\circ{\mathcal{I}}^{-1}=\lim_{n\to\infty}\nu^{{\scriptscriptstyle{({n}})}}\hat T_t\circ{\mathcal{I}}^{-1}=\lim_{n\to\infty}\nu^{{\scriptscriptstyle{({n}})}}T_t\circ{\mathcal{I}}^{-1}=\lim_{n\to\infty}\mu^{{\scriptscriptstyle{({n}})}}P_t,\qquad t>0,$$ concluding the proof.
We continue with the proofs of the results in Section \[sec:densities\].
Let $v=[u]\in{\mathcal{V}}$ and assume that $u$ is not identically $0$ or $1$, since otherwise the statement is trivial. We will show that for each ${{\mathbf x}}\in {\ensuremath{\mathbb{R}}}^{n,\uparrow}$ and $t>0$, we have $$\begin{aligned}
\label{eq:density-variant}
p_{[u]}(t,{{\mathbf x}}) =
\lim_{\epsilon\downarrow0}\frac{1}{(2\epsilon)^n}\, {\ensuremath{\mathbb{P}}}_v\left(\bigcap_{i=1}^n \big\{|{{\mathbf X}}_t\cap [x_i-\epsilon,x_i+\epsilon]|\text{ is odd}\big\}\right).\end{aligned}$$
In order to show , let $(u_t)_{t \geq 0}$ be the $\mathrm{CSVM}_u$-process from Theorem \[thm:voter\]. By the representation of entrance laws in Theorem \[thm:characterization\_entrance\_laws\]b), the $n$-particle density is given by $$p_{[u]}(t,{{\mathbf x}})=\lim_{\epsilon\to0}\frac{1}{(2\epsilon)^n}{\ensuremath{\mathbb{P}}}_{u}\left(\bigcap_{i=1}^n\big\{{\mathcal{I}}(u_t)\cap[x_i-\epsilon,x_i+\epsilon]\neq\emptyset\big\}\right),\qquad {{\mathbf x}}\in {\ensuremath{\mathbb{R}}}^{n,\uparrow},\;t>0.$$
First we argue that for any $x\in{\ensuremath{\mathbb{R}}}$, $$\begin{aligned}
\label{eq:one-interface-only}
\lim_{\epsilon\to0} {\ensuremath{\mathbb{P}}}_{u}\Big( |{\mathcal{I}}(u_t)\cap [x-\epsilon,x+\epsilon]|>1 \;\Big|\; |{\mathcal{I}}(u_t)\cap [x-\epsilon,x+\epsilon]|\geq 1 \Big)=0.\end{aligned}$$ We will use the graphical construction of $u_t(x)=\chi_{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}}}$ via the dual Brownian web, recall . Again we may assume that ${\mathcal{I}}(u_t)\cap\{x-\epsilon,x+\epsilon\}=\emptyset$ since this event has full probability under ${\mathbb{P}}_u$. By the coalescence property, $\widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x-\epsilon}})}}_0 = \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x+\epsilon}})}}_0$ implies that $\widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,y_1}})}}_0=\widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,y_2}})}}_0$ for all $y_1,y_2\in [x-\epsilon,x+\epsilon]$, and therefore ${\mathcal{I}}(u_t)\cap [x-\epsilon,x+\epsilon]=\emptyset$. Therefore the existence of an interface point in $[x-\epsilon,x+\epsilon]$ implies $\widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x-\epsilon}})}}_0\neq \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x+\epsilon}})}}_0$. Similarly, the existence of two interface points in $[x-\epsilon,x+\epsilon]$ implies that there exists $y\in[x-\epsilon,x+\epsilon]$ so that $\widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x-\epsilon}})}}_0\neq \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,y}})}}_0$ and ${\mathcal{W}}^{{\scriptscriptstyle{({t,y}})}}_0\neq {\mathcal{W}}^{{\scriptscriptstyle{({t,x+\epsilon}})}}_0$. As $\epsilon\to0$, the probability that three Brownian motions started in $[x-\epsilon,x+\epsilon]$ do not meet decays faster than the probability that $\widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x-\epsilon}})}}_0\neq \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x+\epsilon}})}}_0$, proving $$\begin{aligned}
\label{eq:one-interface2}
\lim_{\epsilon\to0} {\ensuremath{\mathbb{P}}}_{}\left( |{\mathcal{I}}(u_t)\cap [x-\epsilon,x+\epsilon]|>1 \;\middle|\; \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x-\epsilon}})}}_0\neq \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x+\epsilon}})}}_0 \right)=0.\end{aligned}$$ Further, note that $\{u_t(x-\epsilon)\neq u_t(x+\epsilon)\}$ implies the existence of an interface point in $[x-\epsilon,x+\epsilon]$, thus $$\begin{aligned}
&\frac{{\ensuremath{\mathbb{P}}}_{}\left( |{\mathcal{I}}(u_t)\cap [x-\epsilon,x+\epsilon]|>1 \right)} {{\ensuremath{\mathbb{P}}}_{}\left(|{\mathcal{I}}(u_t)\cap [x-\epsilon,x+\epsilon]|\geq 1 \right)}
\leq \frac{{\ensuremath{\mathbb{P}}}_{}\left( |{\mathcal{I}}(u_t)\cap [x-\epsilon,x+\epsilon]|>1 \right)} {{\ensuremath{\mathbb{P}}}_{}\left( u_t(x-\epsilon)\neq u_t(x+\epsilon) \right)} \\
&= \frac{{\ensuremath{\mathbb{P}}}_{}\left( |{\mathcal{I}}(u_t)\cap [x-\epsilon,x+\epsilon]|>1 \;\middle|\; \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x-\epsilon}})}}_0\neq \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x+\epsilon}})}}_0 \right)} {{\ensuremath{\mathbb{P}}}_{}\left( u_t(x-\epsilon)\neq u_t(x+\epsilon) \;\middle|\; \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x-\epsilon}})}}_0\neq \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x+\epsilon}})}}_0 \right)}.\end{aligned}$$ Together with and the fact that $$\liminf_{\epsilon\to0} {\ensuremath{\mathbb{P}}}\left(u_t(x-\epsilon)\neq u_t(x+\epsilon)\;\middle|\; \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x-\epsilon}})}}_0\neq \widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x+\epsilon}})}}_0 \right)>0,$$ we obtain . But this clearly implies $$\begin{aligned}
p_{[u]}(t,x)&=\lim_{\epsilon\to0}\frac{1}{2\epsilon}\,{\ensuremath{\mathbb{P}}}_{u}\left(\big\{{\mathcal{I}}(u_t)\cap[x-\epsilon,x+\epsilon]\neq\emptyset\big\}\right)\\
&=\lim_{\epsilon\to0}\frac{1}{2\epsilon}\,{\ensuremath{\mathbb{P}}}_{u}\left(\big\{|{\mathcal{I}}(u_t)\cap[x-\epsilon,x+\epsilon]|=1\big\}\right)\end{aligned}$$ and therefore also $$p_{[u]}(t,x)=\lim_{\epsilon\to0}\frac{1}{2\epsilon}\,{\ensuremath{\mathbb{P}}}_{u}\left(\big\{|{\mathcal{I}}(u_t)\cap[x-\epsilon,x+\epsilon]|\text{ is odd}\big\}\right),\qquad x\in{\mathbb{R}},\; t>0,$$ which is for $n=1$. Clearly, at the expense of more notation, this argument can be generalized, proving for arbitrary $n\in{\mathbb{N}}$. Now follows directly from .
[ By Proposition \[prop:density-n\], we know that $$\begin{aligned}
p_{[u]}(t,x) = \lim_{\epsilon\to0} \frac{1}{2\epsilon}\, {\mathbb{P}}\left(\chi_{Y_t^{{\scriptscriptstyle{({x-\epsilon}})}}}\ne\chi_{Y_t^{{\scriptscriptstyle{({x+\epsilon}})}}}\right),\qquad x\in {\ensuremath{\mathbb{R}}},\;t>0. \end{aligned}$$ With $\tau^{{\scriptscriptstyle{({x,\epsilon}})}}:=\inf\{s>0:Y_{s}^{{\scriptscriptstyle{({x-\epsilon}})}}=Y_{s}^{{\scriptscriptstyle{({x+\epsilon}})}}\}$ denoting the coalescence time of the two Brownian motions, we use that the above probability is zero conditioned on $\{\tau^{{\scriptscriptstyle{({x,\epsilon}})}}\leq t\}$. Together with the fact that conditionally on $\{\tau^{{\scriptscriptstyle{({x,\epsilon}})}}>t\}$, the random variables $\chi_{Y_t^{{\scriptscriptstyle{({x\pm\epsilon}})}}}$ are independent and Bernoulli distributed with parameter $u(Y_t^{{\scriptscriptstyle{({x\pm\epsilon}})}})$, we obtain that $$\begin{aligned}
{\mathbb{P}}\left(\chi_{Y_t^{{\scriptscriptstyle{({x-\epsilon}})}}}\ne\chi_{Y_t^{{\scriptscriptstyle{({x+\epsilon}})}}}\right)
&={\mathbb{E}}\left[{\ensuremath{\mathbbm{1}}}_{\{\tau^{{\scriptscriptstyle{({x,\epsilon}})}}>t\}}\,h_u(Y_t^{{\scriptscriptstyle{({x-\epsilon}})}},Y_t^{{\scriptscriptstyle{({x+\epsilon}})}})\right],\end{aligned}$$ where $h_u(\cdot,\cdot)$ is the function defined in . Consequently, $$\begin{aligned}
\label{eq:cond-not-meet}
p_{[u]}(t,x) = \lim_{\epsilon\to0} \frac{1}{2\epsilon}\, {\ensuremath{\mathbb{E}}}\left[h_u(Y_t^{{\scriptscriptstyle{({x-\epsilon}})}},Y_t^{{\scriptscriptstyle{({x+\epsilon}})}})\;\middle|\; \tau^{{\scriptscriptstyle{({x,\epsilon}})}}> t\right]{\ensuremath{\mathbb{P}}}(\tau^{{\scriptscriptstyle{({x,\epsilon}})}}> t).\end{aligned}$$ The distribution of a two-dimensional Brownian motion “started at $(0,0)$” and conditioned to stay in ${\ensuremath{\mathbb{R}}}^{2,\uparrow}$ for the time interval $[0,t]$ is known and has density $$\begin{aligned}
\label{eq:2-non-col-BM}
\frac{y_2-y_1}{2t^{3/2}\sqrt{\pi}}e^{-\frac{|{{\mathbf y}}|^2}{2t}},\qquad {{\mathbf y}}=(y_1,y_2) \in {\ensuremath{\mathbb{R}}}^{2,\uparrow},\end{aligned}$$ see e.g. [@KT03], eq. (2.10). Therefore, as $\epsilon\to0$, the conditional expectation in converges to $$\begin{aligned}
\int_{{\ensuremath{\mathbb{R}}}^{2,\uparrow}} h_u(x+y_1,x+y_2)\frac{y_2-y_1}{2t^{3/2}\sqrt{\pi}}e^{-\frac{|{{\mathbf y}}|^2}{2t}} d{{\mathbf y}}.\end{aligned}$$ Moreover, by the reflection principle and the fact that the difference $Y_t^{{\scriptscriptstyle{({x+\epsilon}})}}-Y_t^{{\scriptscriptstyle{({x-\epsilon}})}}$ is a Brownian motion running at twice the speed, we have $${\ensuremath{\mathbb{P}}}(\tau^{{\scriptscriptstyle{({x,\epsilon}})}}>t) = 1-2{\ensuremath{\mathbb{P}}}_0(B_{2t}>2\epsilon) = \int_{-2\epsilon}^{2\epsilon} \frac{1}{\sqrt{2\pi(2t)}}e^{-\frac{r^2}{4t}}dr = \frac{2\epsilon}{\sqrt{\pi t}} + o(\epsilon).$$ Thus, we obtain from that $$\begin{aligned}
p_{[u]}(t,x) = \frac{1}{2\pi t^2}\int_{{\ensuremath{\mathbb{R}}}^{2,\uparrow}} h_u(x+y_1,x+y_2)(y_2-y_1)e^{-\frac{|{{\mathbf y}}|^2}{2t}} d{{\mathbf y}}.\end{aligned}$$ By symmetry of $h_u$, we also have $$\begin{aligned}
p_{[u]}(t,x) = \frac{1}{2\pi t^2}\int_{{\ensuremath{\mathbb{R}}}^{2,\downarrow}} h_u(x+y_1,x+y_2)|y_2-y_1|e^{-\frac{|{{\mathbf y}}|^2}{2t}} d{{\mathbf y}},\end{aligned}$$ and hence $$\begin{aligned}
p_{[u]}(t,x)
&= \frac{1}{4\pi t^2}\int_{{\ensuremath{\mathbb{R}}}^{2}} h_u(x+y_1,x+y_2)|y_2-y_1|e^{-\frac{|{{\mathbf y}}|^2}{2t}} d{{\mathbf y}}\\
&= \frac{1}{4\pi t^2}\int_{{\ensuremath{\mathbb{R}}}^{2}} 2u(x+y_1)(1-u(x+y_2))|y_2-y_1|e^{-\frac{|{{\mathbf y}}|^2}{2t}} d{{\mathbf y}}.\end{aligned}$$ ]{}
Fix ${{\mathbf x}}\in {\ensuremath{\mathbb{R}}}^{n, \uparrow}$. Consider $u\in{\mathcal{M}}^d_1({\ensuremath{\mathbb{R}}})$. Then we partition ${\mathcal{I}}(u)$ into two disjoint sets ${\mathcal{I}}_1(u)$ and ${\mathcal{I}}_2(u)$ by saying that any $ z \in {\mathcal{I}}(u)$ is also in ${\mathcal{I}}_1(u)$ if $u([z-{\varepsilon},z])(1-u)([z,z+{\varepsilon}])) > 0$ for all sufficiently small ${\varepsilon}> 0$ and setting ${\mathcal{I}}_2(u) := {\mathcal{I}}(u) \setminus {\mathcal{I}}_1(u)$.
Then the thinning procedure corresponds to randomly choosing ${\mathcal{I}}_1(u)$ or ${\mathcal{I}}_2(u)$ with probability $\tfrac12$ each. Similarly to the proof of Proposition \[prop:density-n\] (recall ), we obtain that $$\begin{aligned}
&\lim_{\epsilon\downarrow0} \frac{1}{(2\epsilon)^n}\,{\ensuremath{\mathbb{P}}}_u\left(\bigcap_{i=1}^n \big\{{\mathcal{I}}_1(u_t)\cap [x_i-\epsilon,x_i+\epsilon] \neq \emptyset\big\} \right)\\
&=\lim_{\epsilon\to0}\frac{1}{(2\epsilon)^n}\,{\mathbb{P}}_u\left(\bigcap_{i=1}^n\big\{|{\mathcal{I}}_1(u_t)\cap[x_i-\epsilon,x_i+\epsilon]|=|{\mathcal{I}}(u_t)\cap[x_i-\epsilon,x_i+\epsilon]|=1\big\}\right)\\
&=\lim_{\epsilon\to0}\frac{1}{(2\epsilon)^n}\,{\mathbb{P}}_u\left(\bigcap_{i=1}^n\big\{u_t(x_i-\epsilon)=1,u_t(x_i+\epsilon)=0\big\}\cap\big\{|{\mathcal{I}}(u_t)\cap[x_i-\epsilon,x_i+\epsilon]|=1\big\}\right)\\
&=\lim_{\epsilon\to0}\frac{1}{(2\epsilon)^n}\,{\mathbb{P}}_u\left(\bigcap_{i=1}^n\big\{u_t(x_i-\epsilon)=1,u_t(x_i+\epsilon)=0\big\}\right),\end{aligned}$$ where for the second equality we used that on the event $\big\{|{\mathcal{I}}(u_t)\cap[x_i-\epsilon,x_i+\epsilon]|=1\big\}$ that there is exactly one interface point in $[x_i-\epsilon,x_i+\epsilon]$, this interface point belongs to ${\mathcal{I}}_1(u_t)$ iff $u_t(x_i-\epsilon)=1$ and $u_t(x_i+\epsilon)=0$. Now we use again the graphical construction and argue as in the proof of Thm. \[thm:density\]: With $\tau^{{\scriptscriptstyle{({{{\mathbf x}}, \epsilon}})}}$ denoting the first collision time in the system of coalescing Brownian motions $\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i\pm\epsilon}})}}$, we have $$\begin{aligned}
&{\mathbb{P}}_u\left(\bigcap_{i=1}^n\big\{u_t(x_i-\epsilon)=1,u_t(x_i+\epsilon)=0\big\}\right)={\mathbb{P}}\left(\bigcap_{i=1}^n\big\{\chi_{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i-\epsilon}})}}}=1,\chi_{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i+\epsilon}})}}}=0\big\}\right)\\
&={\mathbb{P}}\left(\bigcap_{i=1}^n\big\{\chi_{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i-\epsilon}})}}}=1,\chi_{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i+\epsilon}})}}}=0\big\}\,\middle|\,\{\tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t\}\right)\,{\mathbb{P}}(\tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t)\\
&={\mathbb{E}}\left[\prod_{i=1}^n u(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i-\epsilon}})}})(1-u(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i+\epsilon}})}}))\,\middle|\, \tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t\right]\,{\mathbb{P}}(\tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t).\end{aligned}$$ and thus $$\begin{aligned}
&\lim_{\epsilon\downarrow0} \frac{1}{(2\epsilon)^n}\,{\ensuremath{\mathbb{P}}}_u\left(\bigcap_{i=1}^n \big\{{\mathcal{I}}_1(u_t)\cap [x_i-\epsilon,x_i+\epsilon] \neq \emptyset\big\} \right)\\
&= \lim_{\epsilon\downarrow0} \frac{1}{(2\epsilon)^n}\,{\mathbb{E}}\left[\prod_{i=1}^n u(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i-\epsilon}})}})(1-u(\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x_i+\epsilon}})}}))\,\middle|\, \tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t\right]\,{\mathbb{P}}(\tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t).\end{aligned}$$
Analogously, $$\begin{aligned}
&\lim_{\epsilon\downarrow0} \frac{1}{(2\epsilon)^n}\,{\ensuremath{\mathbb{P}}}_u\left(\bigcap_{i=1}^n \big\{{\mathcal{I}}_2(u_t)\cap [x_i-\epsilon,x_i+\epsilon] \neq \emptyset\big\} \right)\\
&= \lim_{\epsilon\downarrow0} \frac{1}{(2\epsilon)^n}\,{\ensuremath{\mathbb{E}}}\left[\prod_{i=1}^n (1-u(\widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x_i-\epsilon}})}}_0))u(\widehat{\mathcal{W}}^{{\scriptscriptstyle{({t,x_i+\epsilon}})}}_0)\,\middle|\,\tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t\right]\,{\mathbb{P}}(\tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t).\end{aligned}$$ With $q(t,{{\mathbf x}})=\lim_{\epsilon\to0}\frac{1}{(2\epsilon)^n}\,{\mathbb{P}}(\tau^{{\scriptscriptstyle{({{{\mathbf x}},\epsilon}})}}>t)$, the claim follows by choosing ${\mathcal{I}}_1$ or ${\mathcal{I}}_2$ with probability $\tfrac12$.
Appendix
========
On the construction of annihilating and coalescing Brownian motions {#sec:construction}
-------------------------------------------------------------------
The construction of a *finite* system of aBMs (or cBMs) is of course straightforward: If ${{\mathbf x}}\subseteq{\mathbb{R}}$ is finite, take a collection of independent Brownian motions $\{(B_t^{{\scriptscriptstyle{({x}})}})_{t\ge0}:x\in{{\mathbf x}}\}$ indexed by and starting from ${{\mathbf x}}$, and let them run until the first collision time $$\tau:=\inf\{t>0\,|\, \exists\, y,z\in{{\mathbf x}},\,y\ne z: B_t^{{\scriptscriptstyle{({y}})}}= B_t^{{\scriptscriptstyle{({z}})}}\}>0.$$ Note that the collision pair $(y,z)$ is uniquely defined. At time $\tau$, restart the system with the new initial condition where the collision pair is removed from the configuration. An analogous procedure works for cBMs: [ Here, we do not remove the collision pair, but replace it by (two copies of) a single Brownian motion, reflecting the fact that the colliding particles ‘merge’ and evolve together. ]{}
If the initial condition ${{\mathbf x}}\in{\mathcal{D}}$ is infinite, clearly the above procedure does not work any longer since it may happen that the first collision time $\tau$ equals zero with positive probability. The obvious idea to deal with this problem is to approximate the (discrete, hence countable) set ${{\mathbf x}}=\{x_1,x_2,\ldots\}$ by finite sets ${{\mathbf x}}_{n}:=\{x_1,\ldots,x_n\}$ and to show that as $n\to\infty$, the system of aBMs ${{\mathbf X}}^{{{\mathbf x}}_n}$ converges in a suitable sense to ${{\mathbf X}}^{{\mathbf x}}$. See for example [@TZ11 Sec. 4.1] for a weak convergence approach which works for both aBMs and cBMs.
Another possibility is to define the infinite annihilating system by restriction from the corresponding infinite coalescing system, the latter of which can be constructed by *monotonicity*: For each $n\in{\mathbb{N}}$, let $({{\mathbf Y}}^{{{\mathbf x}}_n}_t)_{t\ge0}$ denote a system of cBMs starting from the finite set ${{\mathbf x}}_n$. It is well-known that there exists a coupling such that almost surely, $${{\mathbf Y}}^{{{\mathbf x}}_n}_t\subseteq{{\mathbf Y}}^{{{\mathbf x}}_{n+1}}_t\qquad\text{for all }t>0,\,n\in{\ensuremath{\mathbb{N}}}.$$ Using this, the infinite coalescing system $({{\mathbf Y}}_t^{{{\mathbf x}}})_{t\ge0}$ can be constructed pathwise as a monotone limit. (Note that this monotonicity property does not hold for aBMs, since adding another annihilating Brownian motion by going from ${{\mathbf x}}_n$ to ${{\mathbf x}}_{n+1}$ might kill a previous one.) Having constructed the infinite coalescing system, define for $y\in{{\mathbf Y}}_t^{{\mathbf x}}$ $$C(t,y):=\#\{x \in {{\mathbf x}}: \text{there exists a coalescing path from }(0,x) \text{ to } (t,y)\}$$ and let $${{\mathbf X}}^{{\mathbf x}}_t:=\{y\in{{\mathbf Y}}^{{\mathbf x}}_t\,|\,C(t,y)\text{ is odd}\}.$$ Then it is easy to see that $C(t,y)$ is almost surely finite and that $({{\mathbf X}}^{{\mathbf x}}_t)_{t\ge0}$ defines a system of aBMs starting from ${{\mathbf x}}$, see [@HOV15 Lemma 5.15]. Moreover, we have ${{\mathbf X}}_t^{{{\mathbf x}}_n}\to{{\mathbf X}}_t^{{\mathbf x}}$ *pathwise* almost surely, although the limit is not monotone w.r.t. set inclusion.
Technical lemmas {#sec:technicalities}
----------------
\[lem:compact\] The space ${\mathcal{M}}_1({\mathbb{R}})$ with the topology of vague convergence is metrizable and compact.
We clearly have ${\mathcal{M}}_1({\mathbb{R}})\subseteq B$, where $B$ denotes the closed unit ball in $L^\infty({\mathbb{R}})$. It is well known that the latter space is (isometrically isomorphic to) the topological dual of $L^1({\mathbb{R}})$, and easy to see that vague convergence on $B$ is equivalent to convergence w.r.t. the weak-$*$-topology of $L^\infty({\mathbb{R}})=L^1({\mathbb{R}})^*$ restricted to $B$. Moreover, it is easy to check that ${\mathcal{M}}_1({\mathbb{R}})$ is vaguely closed in $B$. As a consequence of the Banach-Alaoglu theorem (see e.g. [@DS58 Cor. V.4.3], ${\mathcal{M}}_1({\mathbb{R}})$ is compact in the weak-$*$-topology, hence vaguely compact. Moreover, by [@DS58 Thm. V.5.1] it is also metrizable.
\[lem:dense\] ${\mathcal{M}}_1^d({\mathbb{R}})$ is dense in ${\mathcal{M}}_1({\mathbb{R}})$.
First consider $u(\cdot)\equiv\lambda\in(0,1)$ constant. Define $u^{{\scriptscriptstyle{({n}})}}\in{\mathcal{M}}_1^d({\mathbb{R}})$ by $$u^{{\scriptscriptstyle{({n}})}}:=\sum_{k\in{\mathbb{Z}}}{\ensuremath{\mathbbm{1}}}_{[\frac{k}{n},\frac{k+\lambda}{n}]},$$ so that the interface is a translation invariant lattice ${\mathcal{I}}(u^{{\scriptscriptstyle{({n}})}})=\frac{1}{n}({\mathbb{Z}}+\{0,\lambda\})\in{\mathcal{D}}$. Then we have $$\langle u^{{\scriptscriptstyle{({n}})}},\phi\rangle=\sum_{k\in{\mathbb{Z}}}\int_{\frac{k}{n}}^{\frac{k+\lambda}{n}}\phi(x)\,dx\to\lambda\int_{\mathbb{R}}\phi(x)\,dx=\lambda\langle u,\phi\rangle$$ for all $\phi\in{\mathcal{C}}_c({\mathbb{R}})$, i.e. $u^{{\scriptscriptstyle{({n}})}}\to u$ in the topology of ${\mathcal{M}}_1({\mathbb{R}})$. By an analogous construction on compact intervals $[a,b]\subseteq{\mathbb{R}}$ and linearity, we see that we can approximate any step function $$u=\sum_{j=1}^N\lambda_j{\ensuremath{\mathbbm{1}}}_{[a_j,b_j]},\qquad \lambda_j\in(0,1),\; a_j<b_j$$ by elements of ${\mathcal{M}}_1^d({\mathbb{R}})$. Write ${\mathcal{T}}({\mathbb{R}})$ for the space of all step functions on ${\mathbb{R}}$. Now suppose that $u\in{\mathcal{M}}_1({\mathbb{R}})$ is integrable. Since ${\mathcal{T}}({\mathbb{R}})$ is dense in $L^1({\mathbb{R}})$ (w.r.t. the $L^1$-norm, see e.g. [@LL01 Thm. 1.18]), we conclude that any such $u$ can be approximated by step functions in the topology of vague convergence. Finally, if $u\in{\mathcal{M}}_1({\mathbb{R}})$ is arbitrary, we define $u^{{\scriptscriptstyle{({K}})}}:=u{\ensuremath{\mathbbm{1}}}_{[-K,K]}\in{\mathcal{M}}_1({\mathbb{R}})\cap L^1({\mathbb{R}})$ so that $u^{{\scriptscriptstyle{({K}})}}\to u$ vaguely as $K\to\infty$. We have thus shown that $${\mathcal{M}}_1^d({\mathbb{R}})\subseteq {\mathcal{T}}({\mathbb{R}})\cap{\mathcal{M}}_1({\mathbb{R}})\subseteq L^1({\mathbb{R}})\cap{\mathcal{M}}_1({\mathbb{R}})\subseteq{\mathcal{M}}_1({\mathbb{R}}),$$ where each inclusion is dense w.r.t. the topology of vague convergence on ${\mathcal{M}}_1({\mathbb{R}})$. This establishes the assertion of the lemma.
[**Acknowledgments.**]{} This project received financial support by the German Research Foundation (DFG) within the DFG Priority Programme 1590 ‘Probabilistic Structures in Evolution’, grant OR 310/1-1.
[^1]: Institut für Mathematik, Technische Universität Berlin, Straße des 17. Juni 136, 10623 Berlin, Germany.
[^2]: Department of Mathematical Sciences, University of Bath, Claverton Down, Bath, BA2 7AY, United Kingdom.
[^3]: Fakultät für Mathematik und Informatik, Universität Leipzig, Augustusplatz 10, 04109 Leipzig, Germany.
[^4]: Of course, the density $u\in{\mathcal{M}}_1({\mathbb{R}})$ is only defined up to a Lebesgue-null set, but all our results will be independent of the version of $u$ we choose.
[^5]: In fact, one is tempted to just take a family $(\chi_x)_{x\in{\mathbb{R}}}$ of independent Bernoulli random variables with parameter $u(x)$, respectively, and to put $u_t(x):=\chi_{\widehat{\mathcal{W}}_0^{{\scriptscriptstyle{({t,x}})}}}$. This however does not work since $x\mapsto \chi_x$ is not measurable.
|
Last year, he ranked 158th on the PGA Tour in putting. The sight of him yipping putts in the final pairing with Tiger Woods at last year's British Open led to doubts that he would ever win a major.
But Garcia looked like a new man Thursday, thanks to his new best friend. Armed with a belly putter, he dropped virtually everything en route to a 65 at Carnoustie. At 6-under, he had a two-shot lead over Paul McGinley after the first round of the British Open.
Garcia's mood was decidedly different from the last time he played an opening round at Carnoustie. As a 19-year old in 1999, he walked off in tears after an 89 on a decidedly tougher track.
"Most improved," Garcia said of his 24-stroke gain.
Most improved also applies to his putting. About a month ago, Garcia decided he had to make a drastic change after missing the cut in the U.S. Open.
"There's nothing I hate more than not being able to start the putt on line," Garcia said. "Making putts is great, of course, but at least feel like you're hitting good putts. They're burning the edge and you know you have a chance. That keeps you going.
"When you start hitting the putt and as soon as you hit it you know it doesn't have a chance of going in, it's pretty frustrating."
Garcia had resisted going to the belly putter. Vijay Singh, who uses one, had been urging him to make the switch for more than a year. Garcia's father Victor, who is also his coach, has been using a belly putter for a year.
"He always told me just be open-minded, don't close yourself down, realize there are more options out there," Garcia said.
Garcia said he felt comfortable with the short putter, although he might have been a minority of one in that regard. Finally he decided to experiment with the belly putter, and immediately he saw the results. The beauty of the belly putter is lessening the impact of twitchy hands on the stroke.
"You can't imagine the number of good putts I hit on the front nine that didn't go in," Garcia said. "But all of them looked like they were going in, and that's the beautiful thing about it. That's what I love about putting."
That might be the first time anybody has heard Garcia talk about his love of putting.
Now one round doesn't necessarily count as a trend. He could revert quickly to his old troubles on the greens.
But at 27, Garcia desperately wants to begin reaching the high expectations set for him, and that means winning majors. He bristled when somebody told him that nobody had won a major using a belly putter.
"You shouldn't say that. That's no good," Garcia said. "You guys are always trying to find something. If I play like I played today, maybe that will change soon. It's just about getting the ball in the hole. If I have to use, I don't know, a plastic bag to get in the hole, I'll use whatever."
A plastic bag? If the belly putter doesn't work, that could be next on Garcia's list. * |
1. Field of the Invention
This invention relates generally to image processors and, more particularly, to an image processor capable of correcting shaking of an image.
2. Description of the Related Art
A correlation method based on correlation calculation and a block matching method are known for use as a movement vector detection method that is necessary for an image coding apparatus or an apparatus for correcting image shaking (movement).
In a block matching method, an input image signal is divided with respect to a plurality of blocks of a suitable size (e.g., 8 pixels.times.8 lines), differences from pixels in a certain area of the preceding field (or frame) are calculated with respect to each block, and a block in the preceding field for which the sum of absolute values of the differences is at a minimum is searched for. The relative shift between the related blocks is expressed as a movement vector.
A conventional image shaking correction apparatus which detects a movement vector of an image by using the block matching method and corrects shaking of the image by using the movement vector will be described below with reference to the drawings.
FIG. 1 is a block diagram of a conventional image shaking correction apparatus.
A block 1 represents an A/D converter for converting an input analog image signal into a digital signal.
A block 2 represents a Y/C separation circuit for extracting only a Y (luminance) signal from the digital signal converted from the input image signal.
A block 3 represents a memory for storing the Y signal output from the Y/C separation circuit 2.
A block 4 represents a movement vector detection circuit 4 for detecting the amount of movement of a vector by comparing the Y signal of a preceding frame stored in the memory and the Y signal of the present frame output from the Y/C separation circuit 2.
A block 5 represents a memory read control circuit for determining an image portion to be read from an image stored in the memory 3 by receiving movement vector information from the movement vector detection circuit 4.
A block 6 represents a memory for storing an output signal from the A/D converter 1.
A block 7 represents an interpolation/enlargement circuit for interpolation and enlargement processing of an image signal obtained by reading a portion of the picture stored in the memory 6. Through the circuit 7, an image signal presenting an image enlarged to a predetermined picture size is output.
A block 8 represents a D/A converter for converting a digital signal into an analog signal.
In the conventional image shaking correction apparatus arranged as described above, an image signal representing a shaking image is converted into a digital signal by the A/D converter 1 and distributed to two lines.
Through one of these lines, the digital signal is sent to the Y/C separation circuit 2, and only the Y signal is extracted and input to the memory 3 and the movement vector detection circuit 4.
In the movement vector detection circuit 4, a movement vector is detected by comparing signal data of the present and preceding fields with respect to each of a plurality of blocks in accordance with the block matching method. More specifically, a correlation between the present and preceding fields is calculated with respect to each block, and a block of the preceding field at which the result of correlation calculation, i.e., a correlation value, is at a minimum is searched for. The relative shift of the block thereby found is set as a movement vector.
On the basis of the detected movement vector detected by the movement vector detection circuit 4, the memory read control circuit 5 changes addresses at which the image signal stored in the memory is read out, thereby changing the image position.
The operation of this apparatus will be described in more detail with reference to FIG. 2. Certain frames of an input image, such as frames F1, F2 . . . shown in FIG. 2, are extracted. One of such frames has been translated in an entire picture on the basis of a movement vector A so that shaking of the picture is cancelled out. The position of the extracted frame is shifted by changing the read addresses.
In this case, it is necessary to enlarge the size of the image output from the memory 6, since image shaking correction is performed by extracting certain frames within the normal picture size. Therefore, the image output from the memory 6 undergoes interpolation/enlargement processing in the interpolation/enlargement circuit 7.
The image signal processed for interpolation/enlargement is converted into an analog signal by the D/A converter circuit 8 before it is output.
The conventional movement vector detection processing is performed in the above-described manner. However, it is natural that when a movement vector is detected by the movement vector detection means with respect to each block, the detected movement vector depends largely upon the image state (pattern) of blocks as processing objects.
The influence of image patterns will be described with reference to the drawings.
FIG. 3 shows an ordinary image. FIGS. 4(a) through 4(d) are graphs of variations in correlation values depending upon image states (patterns).
It can be understood that if the above-described block matching is performed with respect to an ordinary image pattern, e.g., a portion a shown in FIG. 3, a strong correlation is exhibited at one point by correlation calculation, as shown in FIG. 4(a), and that a movement vector can be detected with high reliability in such a case.
However, if block matching is performed with respect to a portion b shown in FIG. 3 (e.g., a low-contrast pattern, such as an image of a sky or a white wall), a sufficiently strong correlation cannot be obtained by correlation calculation, as shown in FIG. 4(b), and there is a possibility that a movement vector with a large error will be detected.
Similarly, if block matching is performed with respect to a portion c shown in FIG. 3 (e.g., a reiterative pattern, such as an image of a brick wall or a wire net, consisting of congruent figures), a strong correlation is exhibited at a plurality of points by correlation calculation, as shown in FIG. 4(c), and there is a possibility that a movement vector with a large error will be detected.
Further, if block matching is performed with respect to a portion d shown in FIG. 3 (e.g., an image pattern of a roof or blind slats having a strong correlation in one direction but substantially no correlation in other directions), a strong correlation is exhibited at a plurality of points by correlation calculation, as shown in FIG. 4(d), and there is also a possibility that a movement vector with a large error will be detected. |
oh. the half of lsd dream emulators that are urban are always a joy to wander around to. i can't believe i hadn't played it until a few months ago given how many ppl i know who's taste i trust love it. truly something special.
There is a popular narrative form that could fit desistance or detransition, namely demonic possession + successful exorcism. That needs a more complicated story, however, because there also exist real trans who would be happier with reassignment.
goofysdeadwife69 wrote:i took some great shots of sh 2/3 the last time i played them on this computer. i'll occasionally boot up 3 just to wander around the mall that the game starts in. idk why that never came up for me in the "favorite videogame space" cuz it's definitely one of them. playing those games on pcsx2 w/ the graphic settings cranked up rly brings home how timeless those games are. doesn't really work for the first game cuz of how intentionally lo-fi it is tho.
OK so I played this for twenty minutes or so and it bored the living christ out of me. I was just shooting brown people in Dubai. Does it get better? I'm an intensely lazy pacifist gamer so maybe this isnt the game for me.
it's an imperfect game but it gets better. what you've played so far is basically the tutorial level and it ain't great
if i'm gonna speak to its urbex credentials i can tell you that despite its less than perfect qualities it's got one of the most well realized and frankly awe inspiring undergrounds i've every seen in a game. and yeah i'm talking about its sewers and storm drains. for every inch of prague streets you get to explore in this game there's an equal amount of underground space
Germs: Nerawareta Machi (literally "Germs: The Targeted Town") is a Japanese-exclusive open world adventure game for the PlayStation that was developed and published by KAJ. Taking the role of a reporter who has returned to their hometown to work at a newly opened branch office for their newspaper, players must investigate the town as it undergoes a series of unusual events, most notably the grotesque mutation of its inhabitants. This is accomplished primarily by gathering information, but combat sections involving melee weapons and guns also occur periodically. Like many other games of its genre, Germs features an internal clock that runs six times faster than real time, as well as a day/night cycle, both of which affect the schedules of various functions around time, most notably the times when players can enter businesses and use public transportation. Released to little fanfare in Japan, it remains the only game produced by KAJ to date.
There is a popular narrative form that could fit desistance or detransition, namely demonic possession + successful exorcism. That needs a more complicated story, however, because there also exist real trans who would be happier with reassignment.
i was pretty thoroughly unimpressed with deus ex human revolution as a whole but i think about wandering the streets of detroit and hong kong pretty often.
invisible war is prettttty bad but i have tons of fondness for it just because it was the first game i played that even hinted at this type of potential.
There is a popular narrative form that could fit desistance or detransition, namely demonic possession + successful exorcism. That needs a more complicated story, however, because there also exist real trans who would be happier with reassignment.
that reminds me of a more urban chulip, another game i've always wanted to play but never gotten around to.
There is a popular narrative form that could fit desistance or detransition, namely demonic possession + successful exorcism. That needs a more complicated story, however, because there also exist real trans who would be happier with reassignment.
the last few hours of shenmue ii aren't 'urban'. like you're in a rural forest wandering around with a woman while you talk about animals. but it's sequencing after the first 50-75% of the game is in an urban environment helps make it to be one of the most powerful things in games i can think of & helps a relatively constrained game space feel absolutely immense. sega spent millions of dollars making the most expensive games in history at the time and they're 90% stuff like that. why couldn't that have caught on and more games were just wandering around environments and talking to people.
oh, kowloon walled city is a pretty amazing, dense feeling environment on it's own. what a game.
There is a popular narrative form that could fit desistance or detransition, namely demonic possession + successful exorcism. That needs a more complicated story, however, because there also exist real trans who would be happier with reassignment.
mankind divided is 6 bucks on steam right now which i feel is the right price to wander around in a not great deus ex game will report back to this thread with my impressions of it.
There is a popular narrative form that could fit desistance or detransition, namely demonic possession + successful exorcism. That needs a more complicated story, however, because there also exist real trans who would be happier with reassignment.
i was pretty bummed when after having spent a lot of time making a detailed futuristic montreal (the game was developed here) they were basically told eventually to cut the montreal bits to almost nothing. I'd like to know what its like to see your city replicated in a game even in a futuristic one. I think too if i remember correctly it would have been set in my neighborhood.
I was pretty geeked when i played AC Black Flag and at some point in the modern day bits you get to the top of the building and you can see Hochelaga and downtown and it looks like it should to the point i could more or less point out to where i used to live in Hochelag. Of course its easier to do it accurately when it's just a background and not somewhere you can actually go.
Still i hope one day i can roam around in a game replica of this city. It's like not enough of a famous intl destination to make it be a game setting. I think still the most likely to do it are Ubisoft. Make a watch dogs III set here or something.
the first deus ex is definitely better than any of the games that followed but i still dont know if it would qualify as city feeling for this threads purposes. i feel like the scenery is vaguer and leaves more to the imagination and your brain fills in a lot of the details i guess. i really like the paris level because it's so stressful that i've never explored what feels like 75% of it.
god maybe ill give daggerfall a shot instead. i'd almost 100% have more fun with it if it's not impossible to play on my computer.
There is a popular narrative form that could fit desistance or detransition, namely demonic possession + successful exorcism. That needs a more complicated story, however, because there also exist real trans who would be happier with reassignment. |
Effects of starvation on rat liver mRNA translation products.
The synthesis of rat liver protein and RNA decreases with starvation. It is not yet known whether such decreases are regulated strictly at a transcriptional level, or if post-transcriptional controls are also involved. In this study we investigate the effects of 0, 2, or 4 days starvation on the levels of specific, abundant mRNAs in total and polysomal RNA populations. The mRNAs were analyzed by translation in vitro in mRNA-dependent, cell-free, protein synthesizing systems. The resulting polypeptide products were separated by gel electrophoresis and visualized with fluorography. The amount of albumin translated from both polysomal and total cellular mRNA decreased 20-40% with fasting. In contrast, a specific peptide having a molecular mass of approximately 30 kDa increased two- to three-fold in total cellular RNA with a smaller increase observed in polysomal RNA. These changes were maximal at 2 days of starvation. Since starvation is known to cause alterations in liver metabolism the 30-kDa polypeptide may be related to enzymes or other proteins involved in this homeostatic response. |
1. Field of the Invention
This invention relates to an anchor for securely mounting an article of furniture or the like to a support surface and, more particularly, to an anchor for stabilizing modular office system wall panels against undesired movement resulting from seismic activity.
2. History of the Related Art
Modular office systems provide an excellent means for dividing otherwise large open areas into a plurality of smaller, more functional work spaces. A representative example of such a system is disclosed in U.S. Pat. No. 4,685,255 issued Aug. 11, 1987 to Kelley, et al. Typically, these systems comprise a plurality of panels joined together in an end-to-end or other angular relationship. Often times, work surfaces, storage cabinets and the like are cantilevered from the wall panels through the interengagement of hanger brackets and slotted hanger rails. Such systems are well known. An example is shown in U.S. Pat. No. 4,618,192 also to Kelley and issued Oct. 21, 1986.
The wall panels of the presently known systems are stabilized against undesired movement by being secured to each other and, at times, to a fixed surface such as a wall. While this arrangement is perfectly satisfactory for the vast majority of installations, occasionally it may be desirable to provide enhanced stabilization. For example, certain areas of this country, as well as the rest of the world, are subject to unexpected and potentially violent seismic or earthquake activity. These occurrences have been well documented in the press and elsewhere. Consequently, it would be desirable to have a way for securing articles of furniture or the like, and, more particularly, modular office system wall panels, to a floor surface to restrain the article against undesired movement that might result from unexpected structural or tectonic vibrations of unpredictable magnitude. |
[Cerebellar infarct with complications].
Symptoms of circulatory disorders depend on the volume and localization of the cerebellar infarction. An isolated blood vessel occlusion may have various symptoms due to numerous anastomoses in the cerebellar hemispheres. Dominant symptoms are vertigo, vomiting, nausea and balance disorders. Due to a great number of anastomoses between the three cerebral arteries, if only one of them is occluded, expected symptoms rarely occur, so that vertebral angiography performed in the first 24 hours from the onset of clinical picture, points to a lesion of the cerebellar parenchyma on the basis of radiologic signs of occlusion. Further on, brain computerized tomography solves the diagnostic dilemma, but unfortunately, ischemic cerebellar lesions are often detected only in pathoanatomic findings. Development and course of cerebellar malformation are of great importance for occurrence of symptoms. Cerebellar symptoms often occur in cases of brain-stem lesions. In dorsal hemisection lesions the following symptoms occur: vertigo, vomiting, nystagmus, taste disorders and secretory trophic disorders in the gastrointestinal tract. In cases of spreading to the pyramidal neurons the following symptoms occur: a soft palate and vocal cord paresis, swallowing difficulties and disorders of phonological as well as hemiplegia and sensibility disorders. This is a case of a male patient, 60 years of age. He was admitted to the Neuropsychiatric Department because of high blood pressure, severe headache in the back of the head, vertigo and nausea, without vomiting. He complained about weakness in the left side of the body. The somatic status was normal. In regard to neurologic status he had a slow photomotoric reaction of pupils to light, swallowing difficulties, hoarse voice, decreased pharyngeal reflex, symmetrically fast reflexes, walk on wide basis, in Romberg position unstable, whereas other findings were normal. Psychically average. The reaction to therapy was good. Subjectively the patient felt-well, but swallowing difficulties persisted. On the seventh day from admission his status suddenly aggravated, with severe vertigo, vomiting on several occasions, weakness and high temperature, hypotension, tachycardia, left eyelid ptosis with horizontal nystagmus when looking to the left. Lung radiography revealed bilateral pneumonia; computerized tomography of the brain revealed a great ischemic zone in the left cerebellar hemisphere discretely dislocating the brain mass to contralateral side. Resistant hypotension persisted. Surgical findings of the rectum revealed a dark content. Gastroscopy revealed esophageal candidiasis; mucus bleeding was spontaneous and at touch, whereas in the lower third of esophagus there was a mucous lesion. Haematinised blood was found in the stomach and duodenum. The patient's state continually aggravated and after massive stool blood loss a surgery was performed. A great penetrating ulcer was detected at the back duodenal wall 30 mm of size, visualizing the pancreas. A final surgical procedure was performed at the Clinic for Abdominal Surgery in Novi Sad. The postoperative course was good, but on the 7th postoperative day unexpected fatal outcome occurred due to massive lung embolism. Pathoanatomic examination revealed a postencephalomalacic pseudocyst in the left cerebellar hemisphere. This is a case report of a patient with lateral medullar syndrome associated with complications which can only partly be explained by the basic disease. Pneumonia and a secretory trophic disorder of the gastrointestinal tract are frequent complications that can be life threatening is spite of intensive management. |
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.ignite.internal.processors.cache.distributed;
import java.io.Externalizable;
import java.nio.ByteBuffer;
import org.apache.ignite.IgniteCheckedException;
import org.apache.ignite.IgniteLogger;
import org.apache.ignite.internal.GridDirectTransient;
import org.apache.ignite.internal.processors.cache.GridCacheSharedContext;
import org.apache.ignite.internal.processors.cache.transactions.IgniteTxState;
import org.apache.ignite.internal.processors.cache.transactions.IgniteTxStateAware;
import org.apache.ignite.internal.processors.cache.version.GridCacheVersion;
import org.apache.ignite.internal.util.tostring.GridToStringBuilder;
import org.apache.ignite.internal.util.tostring.GridToStringExclude;
import org.apache.ignite.internal.util.typedef.internal.U;
import org.apache.ignite.plugin.extensions.communication.MessageReader;
import org.apache.ignite.plugin.extensions.communication.MessageWriter;
/**
* Response to prepare request.
*/
public class GridDistributedTxPrepareResponse extends GridDistributedBaseMessage implements IgniteTxStateAware {
/** */
private static final long serialVersionUID = 0L;
/** Error. */
@GridToStringExclude
@GridDirectTransient
private Throwable err;
/** Serialized error. */
private byte[] errBytes;
/** Transient TX state. */
@GridDirectTransient
private IgniteTxState txState;
/** */
private int part;
/** */
protected byte flags;
/**
* Empty constructor (required by {@link Externalizable}).
*/
public GridDistributedTxPrepareResponse() {
/* No-op. */
}
/**
* @param part Partition.
* @param xid Lock or transaction ID.
* @param addDepInfo Deployment info flag.
*/
public GridDistributedTxPrepareResponse(int part, GridCacheVersion xid, boolean addDepInfo) {
super(xid, 0, addDepInfo);
this.part = part;
}
/**
* @param part Partition.
* @param xid Lock or transaction ID.
* @param err Error.
* @param addDepInfo Deployment info flag.
*/
public GridDistributedTxPrepareResponse(int part, GridCacheVersion xid, Throwable err, boolean addDepInfo) {
super(xid, 0, addDepInfo);
this.part = part;
this.err = err;
}
/**
* Sets flag mask.
*
* @param flag Set or clear.
* @param mask Mask.
*/
protected final void setFlag(boolean flag, int mask) {
flags = flag ? (byte)(flags | mask) : (byte)(flags & ~mask);
}
/**
* Reags flag mask.
*
* @param mask Mask to read.
* @return Flag value.
*/
protected final boolean isFlag(int mask) {
return (flags & mask) != 0;
}
/** {@inheritDoc} */
@Override public int partition() {
return part;
}
/** {@inheritDoc} */
@Override public Throwable error() {
return err;
}
/**
* @param err Error to set.
*/
public void error(Throwable err) {
this.err = err;
}
/**
* @return Rollback flag.
*/
public boolean isRollback() {
return err != null;
}
/** {@inheritDoc} */
@Override public IgniteTxState txState() {
return txState;
}
/** {@inheritDoc} */
@Override public void txState(IgniteTxState txState) {
this.txState = txState;
}
/** {@inheritDoc} */
@Override public IgniteLogger messageLogger(GridCacheSharedContext ctx) {
return ctx.txPrepareMessageLogger();
}
/** {@inheritDoc} */
@Override public void prepareMarshal(GridCacheSharedContext ctx) throws IgniteCheckedException {
super.prepareMarshal(ctx);
if (err != null && errBytes == null)
errBytes = U.marshal(ctx, err);
}
/** {@inheritDoc} */
@Override public void finishUnmarshal(GridCacheSharedContext ctx, ClassLoader ldr) throws IgniteCheckedException {
super.finishUnmarshal(ctx, ldr);
if (errBytes != null && err == null)
err = U.unmarshal(ctx, errBytes, U.resolveClassLoader(ldr, ctx.gridConfig()));
}
/** {@inheritDoc} */
@Override public boolean writeTo(ByteBuffer buf, MessageWriter writer) {
writer.setBuffer(buf);
if (!super.writeTo(buf, writer))
return false;
if (!writer.isHeaderWritten()) {
if (!writer.writeHeader(directType(), fieldsCount()))
return false;
writer.onHeaderWritten();
}
switch (writer.state()) {
case 8:
if (!writer.writeByteArray("errBytes", errBytes))
return false;
writer.incrementState();
case 9:
if (!writer.writeByte("flags", flags))
return false;
writer.incrementState();
case 10:
if (!writer.writeInt("part", part))
return false;
writer.incrementState();
}
return true;
}
/** {@inheritDoc} */
@Override public boolean readFrom(ByteBuffer buf, MessageReader reader) {
reader.setBuffer(buf);
if (!reader.beforeMessageRead())
return false;
if (!super.readFrom(buf, reader))
return false;
switch (reader.state()) {
case 8:
errBytes = reader.readByteArray("errBytes");
if (!reader.isLastRead())
return false;
reader.incrementState();
case 9:
flags = reader.readByte("flags");
if (!reader.isLastRead())
return false;
reader.incrementState();
case 10:
part = reader.readInt("part");
if (!reader.isLastRead())
return false;
reader.incrementState();
}
return reader.afterMessageRead(GridDistributedTxPrepareResponse.class);
}
/** {@inheritDoc} */
@Override public short directType() {
return 26;
}
/** {@inheritDoc} */
@Override public byte fieldsCount() {
return 11;
}
/** {@inheritDoc} */
@Override public String toString() {
return GridToStringBuilder.toString(GridDistributedTxPrepareResponse.class, this, "err",
err == null ? "null" : err.toString(), "super", super.toString());
}
}
|
News Main Menu
Penn State astronomers ranked high in scientific impact
July 7, 2009
Penn State astronomers ranked high in scientific impact
University Park, Pa. — Penn State astronomers Peter Meszaros and Donald Schneider are among the scientists whose research has the most scientific impact worldwide, according to ScienceWatch, an organization that monitors performance in basic research.
Meszaros is known for his seminal work in developing, with Martin Rees of Cambridge University, the model that explained gamma-ray bursts as titanic cosmic explosions. Meszaros, who won the American Astronomical Society's Rossi Prize in High Energy Astrophysics in 2000, is extensively involved with the Swift satellite, whose Mission Operations Center is run by Penn State. The Swift satellite was launched in 2004 and has detected hundreds of gamma-ray bursts, resulting in a breathtaking number of discoveries about our universe. "This field of study has really blossomed in the past two decades," said Meszaros. "When I arrived at Penn State in 1983, our understanding of gamma-ray bursts was so limited that they could have been located anywhere between the outer solar system to the most distant reaches of the universe. Now we frequently measure the energies and distances from Earth of individual bursts. One burst was so bright that, although it was several billion light years distant, it was visible to the naked eye for a minute."
Schneider has devoted much of his research effort for the past twenty years to the Sloan Digital Sky Survey (SDSS), a multi-institutional project to map a large fraction of the sky. Schneider is the leader of SDSS quasar science research, which to date has identified over 100,000 quasars, although fewer than 10,000 were known when the project began making observations. The SDSS project's quasar research has resulted in a series of discoveries that repeatedly have broken the record for the most distant object yet discovered.
Department Head Lawrence Ramsey noted that he was pleased but not surprised by Science Watch's announcements. "We expect our faculty to be internationally recognized leaders in their fields, and professors Meszaros and Schneider are among those in the department whose research achievements are at this high level," he said. |
As you know from reading this site, I just moved back to America a little over a week ago. Now I am preparing to file for my wives Petition for Alien Relative, or better known as Spousal Visa.
To file you need to send an I-130 form “Petition for Alien Relative” and an I-325a “Biographical Data” form for both your wife and yourself.
Once the petition is approved, you will need to file a G-864 Affidavit of Support to...
For anyone who has been here in the Philippines for a while and still on a tourist visa you know about how the fees at Bureau of Immigration are always changing. When I first moved here, if you waited for your visa extension to be processed, you were charged a 500-peso “Express Line Fee.” This is something they never asked you if you wanted, they just charged you.
Then they made the Express Line fee mandatory. Recently...
In The U.S. just as here in the Philippines, there are different ways of obtaining residency. Here you can receive residency from a marriage to a Filipino, by signing up for a Retirement Visa, there is even a new way if you own a business with ten or more employees.
For a 13a Visa, which you receive after marriage, you apply with your wife. Most of the application is from the Filipina including a petition for your permanent...
Saturday Elena and I joined her group of expat wives, a few of their kids and two other of us husbands and went to use the pool at the Marco Polo Hotel. At one point there was a commotion and I asked what was going on. They all mentioned Noynoy Aquino was there on the other side of the Pavilion Restaurant.
As a Foreigner, I do not write or comment on government issues, politicians or campaigns. It is not the place of...
My Friend and fellow Blogger John Ray of Palawan Anecdotes sent me this article. He has written a few this past week about his beloved past President Corazon Aquino. I hope you will enjoy his words and sentiments about his country which he loves and cares about. These are John Rays words and he asked me to post here for him. Being a guest in his country, I will not comment if I agree or disagree whith his feelings since it...
Many foreigners move to the Philippines for different reasons. Most are men and most are in retirement age. Maybe you just got tired of life and stress in your home country. To move here you must have some savings or pension to support yourself and your possible new wife if you’re single. Maybe you know of a lady from the internet or from a past visit. Maybe you are unattached and think; once you’re here you will find... |
Are you looking the best price for Seagate Barracuda 1TB 7200.11 Internal 3.5" Serial ATA-300 7,200 RPM 32MB Hard Drive - ST31000340AS and you want to get great deals & best buy this product? Or you want to find some reviews from people who have used this product before you buy it?
You are at the right place. I'll give you some tips before you decide to buy Seagate Barracuda 1TB 7200.11 Internal 3.5" Serial ATA-300 7,200 RPM 32MB Hard Drive - ST31000340AS. So you are not disappointed after buying it and in accordance with the product you need.
You can also find price comparisons from sellers, so you really get the best price currently for the Seagate Barracuda 1TB 7200.11 Internal 3.5" Serial ATA-300 7,200 RPM 32MB Hard Drive - ST31000340AS. |
Earlier scanning straight into this informative article, recognize the fact that remaining positive your assume is often cruel in addition to hardly ever joy. In spite of this, it is a requisite with the aim of ought not ensued kept away from if single poverties proper as well as glad years. Luckily, you do not need near method the idea in different utmost frame. You recently need to conduct some time out of your generation and also have a quantity of strength of will. This may well possibly ensue joy. If you want to acquire healthy on the contrary tend not to receive riches meant for a physical education building link or perhaps believe effect gear, tend not to worry. You will discover lots of keeps fit you're able to do without necessitating any involving this specific. You possibly can rush, totter, looks after remains or even push-ups. Never allow an absence of money enter your path of appearing moreover being of the opinion immense.
+
+
Wish the perfect epoch associated with sunlight hours for ones human body to be able to physical exercise. A morning person will find it relatively all to easy to fit into their particular work out normal in advance inside the date, whereas a star that strokes in their utmost down the line within the period ought to remain until finally the morning or else evening in order to train. In the event you figure out whenever your bulk with be offended subsists feeling wearing first-class prepare, you might acquire the most effective products on the cards. Allot ones problems over numerous little sessions throughout the day. With your hard-working program, maybe you have problems establishing tangent 1 hour toward implementation. Nevertheless, should you better to facilitate happy in to minor conferences over the calendar day, you'll contract the identical charity performance as a single stretched meeting. Smooth jobs around the put up is usually turn into in a workout usual. A very worthy method to develop tally ends up being not later than join within a round division. On the whole exercise room commonly suggest sequence group of students and they are an incredible solution to get in affect after that go through populaces. Trainers strength of character impetus a person as a consequence they as a rule star as good song that makes the routine more pleasurable. Put into effect in your own home. Desire near piece of work in your house you will be further likely to follow-through in your own keep fit standard. You may not have to uncultivated point in time guide to the leisure center, parking, or perhaps kill time about the products you need to exhausted. Designing outdated at home will even recover a person funds. Essaying growths the fresh air on the wits. Studies control provided evidence with the intention of adding an problem system in order to your own daily schedule pray lower the chance connected with getting dementia appearing in nearly 60% participating in adult adults. Piece of work launches proteins of which fortifies the brain's neurons with cells which can be straight cooperate for you to memory plus gather. The obvious way to ensure an individual pole with finding standard work out is always to sees to features people take pleasure in executing. Becoming a simple yet effective work out won't have to help mean designing on boring to death devices like treadmills. In its place, observe a thing you love to accomplish like seaming a dance style or even bother a cycle. Repealing masses is usually an superb approach to ended up being robust. In the event you hold stolen burdens or, a person really should inception banned with a little gentle, for example a individual batter dumbbell. With the dumbbell, perform a number of 10 uplifts otherwise reps. If you can elate the dumbbell especially quickly, then anyone really should go up to and including heavier consequence and soon you obtain one who bestows enough amount of resistance without having to be as well minder. |
Related Tags:
Roger Maris is the single-season home run champion, regardless of how many home runs Mark McGwire, Barry Bonds and Sammy Sosa managed to hit. He will be until someone deposits No. 62 over a fence and then does something even more difficult by convincing us that it was done without the benefit of modern chemistry.
That someone could be Chris Davis of the Baltimore Orioles. He has 45 home runs with 38 games left, and is on a pace to threaten the record if not break it.
It won’t be easy, and the pressure will mount with each at-bat. Maris himself started losing hair in clumps as he chased the mark of 60 held by Babe Ruth a half century ago, enduring hate mail and death threats along the way.
At least Davis has some control over what happens on the field. He doesn’t when it comes to the court of public opinion.
If he hits a lot of home runs he must be juiced. Period, end of story. That’s how we look at sluggers now, because what we saw before was for the most part a big lie.
It may not be fair to Davis, but that’s the world we live. Presumed guilty until proven innocent, and don’t bother with the protests of that innocence. We’ve heard them before so many times, that even the lies of Alex Rodriguez and Ryan Braun can’t shock us anymore.
Never mind that Davis has always shown great power and has never, as far as we know, tested positive for anything that might increase that power. Never mind that he himself considers the single season record to be 61, or that he has handled the inevitable questions about where his power comes from directly and without the feigned outrage we’ve seen from others.
“I’ve got nothing to hide,” Davis said at the All-Star game last month. “I want people to know that. I want people to feel like they can get behind me.”
If only it was that easy. If only we could all find a way to believe once again.
It would be a joyful way to finish off a season, must-see TV every night for any baseball fan. It would also be a perfect antidote to a season when the biggest drug scandal went down, and a year when some of the greatest players of their time were denied entry into the Hall of Fame.
But we remember Rafael Palmeiro shaking his finger before Congress and declaring he had never taken steroids. We’ve read transcripts of Bonds saying he believed the clear and the cream were just flaxseed oil. We fell for Sosa and McGwire and their fabricated summer of peace, love and home runs.
Believe that someone clean can break the home run record? Sure, and don’t forget to make sure you leave the Easter Bunny a snack before heading to the ballpark.
If it’s any consolation to Davis, he’s not the only one we find ourselves wondering about. Miguel Cabrera became the first player in 45 years to win the triple crown last year and is in the running for it again this year. He does things so astonishing that Detroit manager Jim Leyland says he is the best right-handed hitter he has ever seen (Bonds was listed as Leyland’s best lefty).
No one has suggested he is dirty, and there’s no evidence that he is. But do you want to risk $99.95 buying his jersey and your time cheering him on when most of the sluggers before him were all putting on a charade?
And then there’s Albert Pujols, who has largely stayed out of the performance enhancing drugs discussion despite his dominant numbers over the last decade. Pujols says he plans to take legal action against former Cardinal Jack Clark for saying on a radio show that Pujols had taken PEDs.
Let’s hope Pujols follows through on his threat to sue. Assuming he’s always been clean he has nothing to lose in a lawsuit and everything to gain. His reputation would be upheld, his status as one of the game’s greats would be cemented, and he can go back to worrying about how to live up to the next eight years on his contract with the Los Angeles Angels.
But lawsuits can be a tricky thing. Does Pujols want an opposing attorney dissecting everything he’s done in his career and everything he’s ever put in his body? Is he so sure of the rightness of his cause that he will risk the kind of scrutiny he’s never seen?
Again, it’s not fair. Pujols shouldn’t have to be defending himself simply because he’s a big hitter. It’s guilt by association because, hey, everyone must be guilty.
That includes Davis, who has already hit 12 more home runs this year than he did all of last season. The closer he gets to the record the more speculation there will be, pressure that Maris didn’t have to deal with when he broke the mark Ruth set in 1927.
It’s not the best time to be chasing one of the game’s most hallowed records. But there will be a time when everyone will have to move on or just give up on the game itself.
About all we can do now is hope that Davis is one of the rare ones who can actually be believed. |
Eastern Tennessee family loses a home, gains a community
In early September, the Stutte family was putting their lives together after their home in Vonore, Tenn. burned down, save a single wall on which an anti-gay slur was spray-painted. But when Carol Ann Stutte did something seemingly innocuous—she cancelled her hair appointment, mentioning the fire—she unlocked the door to a community that she never knew existed in eastern Tennessee.
Her hairdresser, Jaime Combs, and Combs’ partner, Carla Lewis, happened to be active in the eastern Tennessee LGBT community. When thinking of how to help, Lewis thought of PFLAG Marvyille and posted the news on the group’s Facebook page, as news of the fire hadn’t yet surfaced in the media.
Becky Lucas, the teacher and mother who recently formed PFLAG Marvyille and is part of the NIOT network, catapulted into action. Lucas called Carol Ann and asked, “How can we help? What do you need?”
Recalling the phone conversation, Carol Ann said, “At that point, we were still in shock. I just started crying, and said, ‘I don’t know.’”
Their needs ranged from cotton swabs to a place to live. Lucas galvanized more than 60 people who attended the first PFLAG Maryville meeting just days before and asked them to also send out pleas to help the Stuttes. Lewis, of the Eastern Tennessee Equality Council and the Tennessee Transgender Political Coalition, was integral in garnering media coverage and setting up an online donation system.
The media began calling and donations flooded in. Various LGBT organizations and churches stepped forward. Lucas spent hours on the phone speaking with people offering items that ran the gamut from toiletries to their late mother’s furniture. A donated storage unit was filled “five or six times over,” Lucas said.
Of the hundreds of calls Lucas received, only three wished them harm. Though support was overwhelmingly positive, there was the possibility that they could also be targets. Lucas’ 12-year-old asked, “Is someone going to burn down our house?”
Despite the risk, Lucas said, “How I feel about the Stuttes, it’s what I would do for my own son.”
Lucas and Carol Ann spoke daily. The community organized potlucks, welcoming the Stuttes with homecooked meals and a group of supporters. They also assisted in securing the Stuttes a civil rights attorney.
Of Lucas, Carol Ann said, “She literally took the bull by the horns, getting everything we needed. It was like ground zero triage.”
Within weeks, the community raised more than $10,000 for the Stuttes. Monetary donations came from 36 states, the District of Columbia, Australia, Canada, Germany and the United Kingdom.
Carol Ann, her partner Laura and daughter Kimberly are currently living in Knoxville, Tenn. thanks to community support. Carol Ann said the county, the state, and now the FBI is investigating the fire. The Monroe County Sheriff’s Department did not return calls for comment.
The fire did not surprise the Stuttes. Since they moved to Vonore five years ago they had been intimidated multiple times by a neighbor. Along with their home, their lives had been threatened and two of their dogs had been poisoned, Carol Ann said.
“To move to a place like this, where hate is like a disease, and then to go to a place like Maryville and Knoxville, it’s like a family welcoming you with open arms,” Carol Ann said.
The story of the Stuttes is really a story of how a small community came together to make a big difference, paving the way for change in eastern Tennessee.
“I’d say we have a bad reputation for hate crimes in this area. There’s just too many times when something like that happens and it usually gets swept under the rug,” Lewis said. “I want the Stuttes to know … that people like me are not going to let them suffer, we’re going to suffer with them and we’re not going to let this happen again.”
Both Lewis and her partner Combs’ activism was sparked two years ago by another tragedy in eastern Tennessee. Both were present when an armed man opened fire on the congregation at the Tennessee Valley Unitarian Universalist Church in Knoxville, Tenn. in 2008. The church is among six others that have supported the Stuttes.
“It was like feeling like you were all alone and then coming home,” Carol Ann said. “We finally felt safe and protected for the first time in five years.” |
Writing on the Wall for Egyptian/Israeli Peace
Last week, a group of terrorists in Sinai killed 16 Egyptian soldiers before launching a failed attack into Israel. And a few days later, the new Egyptian president, Mohammed Morsi, removed the chief of the armed forces and defense minister, Mohammed Tantawi, along with the army, navy and air force service heads. On the same day, he also cancelled the constitutional addendum restricting presidential powers that Tantawi and the Supreme Council of the Armed Forces had imposed last June. These events tell us much about what lies ahead.
Morsi, the Muslim Brotherhood candidate who won presidential elections, has full executive and legislative authority. He can convene a new constituent assembly to draft a new constitution, without the oversight of the military establishment that has ruled Egypt for six decades.
This means an Islamist constitution. The Brotherhood, the “the mother of all Islamist movements” as Shadi Hamid, a Middle East expert at the Brookings Institution’s Doha Center, puts it, an Islamist organization dating back to 1928, whose leading ideologues, notably Sayyid Qutb, were the precursor of al-Qaeda, will create an Islamist order in Egypt.
The Brotherhood is vehemently anti-American, so expect a slow demise in the alliance into which America poured $60 billion over three decades. Its leader, Muhammad Badi’ said in October 2010 that, “The U.S. is now experiencing the beginning of its end, and is heading towards its demise.”
The Brotherhood is also virulently opposed to Israel’s existence and calls for the rescission of the Egyptian/Israeli peace treaty. Its deputy leader, Rashad al-Bayoumi has described Israel as “enemy entity” and asserted that the existing peace treaty “isn’t binding at all.” Expect Israeli/Egyptian relations – frosty at the best of times – to petrify.
Morsi no longer speaks for the Brotherhood – he resigned on becoming president – but he needn’t: it speaks for itself. And its reaction to the recent terrorist attack, which the Israelis narrowly averted, was to blame it on the Israeli intelligence service “Mossad, which has been seeking to abort the Egyptian revolution.” Hamas, the Palestinian off-shoot of the Brotherhood which controls Gaza and calls in its Charter for the worldwide murder of Jews, took the same line. Its prime minister, Ismail Haniyeh, stated, “The crime itself and what preceded it confirms Israel’s involvement in one way or another.”
Is Mr. Mandel claiming that Mossad, has not been seeking to abort the Egyptian revolution? The Israeli press sure has. For weeks they have been lamenting that they can't prop up the gelatinous and wrinkled form of Hosni Mubarak to rule for just a few more decades.
Common sense: Egypt just had a revolution and now they have a new President. Right after his victory, they kill 16 of their own people, steal a personnel carrier and charge into the Sinai "with President Morsi's help?" This is a strategy? Hogwash.
http://www.facebook.com/introvert321 Shmuel Malov
The only problem Morsi might find with what happened is that those terrorists got ahead of themselves and attacked too early. It is doubtful that he was against it.
EthanP
You never give up do you!
PaulRevereNow
You don't take common sense far enough. Common sense: Islamists kill their own people all the time. Just look at whats going on in Syria. So why do you feign surprise that the MB has killed 16 Egyptian soldiers? The article also stated that Morsi has purged the Egyptian military, firing Hussein Tantawi and others, who posed a threat to his authority. Strategy will follow. But when war breaks out, you will say, "This is a war?"
Advocatus
Come on now, Schlomo: on another thread you've just berated Raymond Ibrahim for referring to unsubstantiated accounts of Egypt's Muslim Brotherhood having crificified some opponents.
Because their war is not with Israel, it is with the Jews. Israel is just the lightning rod.
SKIP
Because their book says it must be so and not just Israelis, this fatwa goes for all non muslims. As one can easily see, muslims take as much pleasure in killing each other as they do killing non muslims.
jacob
REASONS ??????
Just read the "Holy" Koran, which ALL MUSLIMS must obey blindly and you'll find the reasons….
As to the Egyptian-Israeli "Treaty", was it anything but a plain "CEASE FIRE" at best ????
And now with the Muslim Brotherhood in full power, what is there in store for Israel ????
I'm afraid ARMAGEDDON is looming in the horizon….
Kufar Dawg
Bukhari:V4B52N177
"Allah's Apostle said, ‘You Muslims will fight the Jews till some of them hide behind stones. The stones will betray them saying, "O Abdullah (slave of Allah)! There is a Jew hiding behind me; so kill him."'"
Ishaq:441
"Allah's Apostle said, ‘The Hour will not be established until you fight with the Jews, and the stone behind which a Jew will be hiding will say. "O Muslim! There is a Jew hiding behind me, so kill him."'"
In Sahih Muslim: Book 041, Number 6985:
Abu Huraira reported Allah’s Messenger (may peace be upon him) as saying: The last hour would not come unless the Muslims will fight against the Jews and the Muslims would kill them until the Jews would hide themselves behind a stone or a tree and a stone or a tree would say: Muslim, or the servant of Allah, there is a Jew behind me; come and kill him; but the tree Gharqad would not say, for it is the tree of the Jews.
watsa46
They refused to acknowledge Mohamed as …. of "god".
Mburu
"On the same day, ELECTED President Morsi also cancelled the constitutional addendum restricting presidential powers that Tantawi and the UNELECTED Supreme Council of the Armed Forces had ILLEGALLY imposed last June.
EthanP
Well said.
EthanP
What no one mentions is that 10 days ago they were saying Morsi had no power, among other negative things. Yet he dismissed the High Command with not a wimper. As to a conflict with Israel. While I have (IMHO) no doubt that Israels destruction is his, as it is for all Islamists, his ultimate goal, he is no fool. Despite all of the M-1 tanks and F-16s the US has given Egypt, there is no question that the qualitative edge in manpower, the true decisive factor in military power, is firmly with Israel. So the MB in the form of Hamas may fire missles and resort to a continuing campaign of terror, but they know beyond doubt the result of a military confrontation with Israel.
Andy
Israel needs not concern itself with Egypt. If they stick to the agreement fine. If they want war, Israel can rout that failed country in a long weekend. What Israel needs to worry about are its own lefties and mentally challenged with the same mentalities some of their forefathers had in Europe in the 1930s who kept saying that Hitler was just posturing, joking. These and their suicide-voter cousins in America who support Obama.
Marty
egypt is a failed state regardless of what ruler is misgoverning it and destroying the economy. It has little to no chance of becoming democratic since every political movement and ideology refuses to disendorse corruption and hatred of Israel as well as its copt minority. Nothing useful comes out of egypt. The place hasn't progressed since islam wrecked the society 1300 years and burned down the library in Alexandria. islam enthusiastically destroys any society it touches and forces it into a primitive mindset.
Stan Lee
Egypt has its internal problems, religious, organizational. governmental, military. One of the traditional methods of forcing unity in a country is by way of nationalism….by way of war.
Kufar Dawg
Or by blaming the Jews…kinda like MuhamMAD did, Hitler did and most every islamic leader the world has ever seen.
flowerknife_us
Israel needs to throw these problems back in the faces of the Egyptians. It is for Egypt to solve as the Treaty stands. Altering the Treaty any more than it has on the part of the Egyptians is unacceptable. It should not be viewed as an act of peace and good will. Force the World body politic to face who is actually at fault.
Both sides are heavily equipped with American arms. Will the difference between who may actually get re-supplied be as simple as who our next President will be???
Drakken
Egpyt is going back to the dark ages and with 80 million mouths to feed, they will be fed a daily diet of jihad and make no doubt, the muzzys will attack Israel at some point and when they do I hope Israel takes the gloves off and takes the opportunity to push that rats nest called Gaza into the sea and takes the Sinai, then pushes on to Cairo and put all the Imams at their feet where they belong.
Raymond in DC
Egypt's request for 750 troops to be allowed into Sinai to deal with weapons smuggling was just the first of such requests. Not surprisingly, 750 was not enough, so Egypt asked for 750 more. Israel said OK. Funny, things didn't really improve. So now it's a couple of brigades, armored vehicles, gunships and a few fighter jets. Israel hesitates to say No. Now I'm seeing reports of reinforcements *without* Israel's approval.
The right course of action from the beginning, as now, is to demand that if Egypt is granted more rights in Sinai, so too should Israel. Start with the right to conduct aerial reconnaissance, and the right of "hot pursuit". If remilitarization is permitted, all that will be left of the peace treaty would be Israel's right of free passage through the Suez Canal and Straits of Tiran. For that they gave up the Sinai? And how long until those rights too will be circumscribed?
Israel will learn and exercise exponentially the same brutality with which it is faced. That will ensure the 'peace'. When Arabs can no longer wage war and terrorism against Israel with impunity the complexion of this conflict will dramatically change. |
Peripheral type benzodiazepine binding sites are a sensitive indirect index of neuronal damage.
The effects of excitotoxic lesions on the neuronal marker enzymes choline acetyltransferase and glutamate decarboxylase and on the levels of 'peripheral type' benzodiazepine binding sites (PTBBS) (a putative glial marker) have been compared to see whether PTBBS provide a suitable if indirect quantitative index of neuronal damage. Intrastriatal injection of excitotoxic compounds provoked a dose-dependent increase in the levels of PTBBS. The potency order was the following: kainate greater than AMPA greater than N-methyl-D-aspartate (NMDA) greater than quisqualate. The maximal increases in this parameter were 400, 470, 320 and 210% for kainate (12 nmol), AMPA (100 nmol), NMDA (500 nmol) and quisqualate (250 nmol), respectively. 2-Amino-5-phosphonovalerate (100 nmol)--an antagonist of the NMDA receptor subtype--completely blocked the increase in PTBBS induced by NMDA (250 nmol), but was without effect against the other excitotoxins. Increases in binding levels were in general mirrored by a decrease in choline acetyltransferase and glutamate decarboxylase activity. However, PTBBS were a more sensitive indirect index of neuronal damage than neuronal enzymes because the alterations in binding were statistically significant at doses of excitotoxins lower than those causing a loss of marker enzymes. It is concluded that PTBBS are a suitable and sensitive means of detecting discrete neurotoxic changes and that its measurement will help in the study of other pathological and experimental models. |
Modi doing best for Gujarat as PM, don’t miss him: Amit Shah
BJP President Amit Shah on Sunday stated that Narendra Modi was better placed to serve Gujarat as Prime Minister than as Chief Minister and appealed the people not to miss him, while citing statistics to indicate the state had developed during all consecutive BJP governments since 1995.
“There is no need to miss Narendra Modi in Gujarat. He placed Gujarat on the growth map as the Chief Minister, as the Prime Minister he is doing much for the development of the state. Who is more powerful, the Chief Minister or the Prime Minister?” said Shah as he answered questions during a town-hall interaction with youth through multiple digital platforms and video-conferencing across 312 locations.
The event was called Adikham Gujarat (Resolute Gujarat) where the BJP chief claimed to have addressed one lakh youth by taking questions through the social media.
Quipping that “today I am not a politician, but a professor”, Shah claimed 4.7 lakh questions had come and this spoke of the expectations and hopes from the Bharatiya Janata Party (BJP).
Amid applause at the jam-packed Deendayal Hall, a questioner asked: “Politically we feel the absence of Narendra Modi and the scenario does not look the same after he has gone three years ago. What does the party intend to do?”
A grinning Shah said: “As the Prime Minister, he is helping in Gujarat’s development by taking decisions at the snap of a finger. Within 10 days of assuming power in Delhi, the permission to increase the height of the Narmada dam came. And now he has ensured the radial gates are put at the dam.
“He travels across India with the interests of Gujarat in his mind and heart. Modi is very much here.”
Shah claimed there was a sea of difference in Gujarat before 1995, when the first BJP government came to power, and now, when the state’s development numbers had zoomed up. “Don’t ask questions, look at the statistics. It’s all there for everyone to see. All the statistics will be put up on the BJP’s website for the youth to check.”
Asked about the pathetic conditions of the state’s roads, he admitted it was so after the rains but credited the BJP governments for creating the largest road infrastructure in Gujarat. “Many people are today taking a dig at us for the quality of roads but it was during our rule that a wide network of roads was created. We cannot resurface the roads during monsoons and the work on improving roads will happen betweenASeptember 15 and October 2,” he said.
The BJP chief also fielded questions on the contentious issues of GST, flogging of Dalit youngsters in Una and the Hardik Patel agitation for reservations to Patidars.
“GST, the concept of one nation one tax, is a progressive taxation regime. The Congress attempted to bring it in, but we succeeded. I admit there are problems in its implementation but the government is aware of this and making efforts to stabilize to new system in the next five to six months,” Shah said.
He admitted the Una incident was shocking but said this was not the order of the day in Gujarat. “Look at the statistics, the cases of atrocities against Dalits are the lowest in Gujarat. Una was a condemnable and our government has initiated strict legal action against the perpetrators,” he said.
On the Patel quota issue, he referred to the earlier position of the state government about the constitutional and legal challenges of a 50 per cent cap on reservation. “At best, the list of communities included in OBC category can be expanded but a legal process must be followed,” he said, but added that, “All such agitations are generally political motivated ones and end up in a whimper.”
The youngsters in the state were asked to send questions through Twitter, Facebook or by giving a missed call on 7878182182 and register themselves to send questions. Alternatively, they could ask questions on www.adikhamgujarat.com. |
[Compliance with chronic pain treatment: study of demographic, therapeutic and psychosocial variables].
The aims of this study were to identify the prevalence of compliance with drug therapy in patients with chronic pain and analyze the relationships between compliance and characteristics of drug therapy and psychosocial factors (beliefs regarding pain, health locus of control and depression). Thirty patients were evaluated 5 times over a period of 6 months. Total compliance occurred in 43.3% to 56.7% of the patients. Partial compliance and non-compliance were high (40.0%-56.7%). The index of compliance did not vary over the six months. Low compliance related to occurrence of side effects and beliefs that the control of health depended on the patient, pain is a disability, that pain means the presence of physical injury and solicitous behavior of others is desirable when there is manifestation of pain. Knowing the factors involved in compliance enables us to test interventions that optimize it. |
NEWS
Highstreet-killer ASOS reports rapid growth
Online fashion retailer ASOS has reported an 18% increase in pre-tax profits in the six months to February 29th, rising to £21m (€26m) on the back of a 24% increase in sales on a constant currency basis to £648.6m.
The London-based company enjoyed its most lucrative festive period on record, at the height of business on the 'discount holidays' of Black Friday and Cyber Monday the site processed up to nine orders per second.
Overall ASOS increased its sales by 21% to £667m in the UK, US and Europe.
After three years trying to tap into the Chinese market ASOS closed its loss-making business there. The division haemorrhaged money, racking up losses of £2.7m during the six months as it failed to gain a foothold competing against local firms such as e-commerce giant Alibaba.
Chief executive Nick Beighton commented on the company's struggles in China, "There are always challenges as a start-up in a country, but there are additional challenges to being a start-up in China," he said.
ASOS has invested heavily in developing its online store and mobile app which has built up a user base of 1.4 billion and now accounts for almost half of the company's sales. |
Q:
jQuery: Checking for `nodeType == 3` in a function
Expanding Using .text() to retrieve only text not nested in child tags -- how do I turn the answer about checking for nodeType == 3 into a function?
function hasText(element) {
element.contents().filter(function () {
if (element.contents().nodeType == 3) {
return true
}
});
}
$('p').each(function () {
if(hasText($(this))) {
$(this).css('border', '5px solid blue');
}
});
<!-- Should return true -->
<p>Hello</p>
<p>Hello <a>there</a></p>
<!-- Should return false -->
<p><a>Hello</a></p>
<p><a href="#"><img src="#"></a>
A:
You need to return a value. Also, Array.prototype.every might be more appropriate than filter.
function hasText(element) {
return Array.prototype.every.call(element.contents(), function (c) {
return c.nodeType === 3;
});
}
$('p').each(function () {
if (hasText($(this))) {
$(this).css('border', '5px solid blue');
}
});
<script src="https://ajax.googleapis.com/ajax/libs/jquery/2.1.1/jquery.min.js"></script>
<!-- Should return true -->
<p>Hello</p>
<!-- Should return false -->
<p><a>Hello</a></p>
|
Training and Employment
The Dallas County Veteran Services offers employment services to Texas veterans and helps employers find qualified veteran job applicants. Our vision is that every veteran and eligible veteran spouse in the State of Texas have access to long term and meaningful work.
Texas Workforce Commission(TWC)
The Texas Workforce Commission (TWC) will be the primary entity used by Veterans Referral and Resource Specialists (VRRSs) to assist returning veterans as they transition to civilian life.
This program is coordinated by Department of Labor Veterans Employment and Training DOL VETS on a national level to help create a seamless, personalized assistance network to ensure that seriously wounded and injured service members who cannot return to active duty are trained for rewarding careers in the public or private sector. REALifelines provides immediate intervention and casework support for veterans and their families. RLL supports the economic recovery and employment of transitioning service members and their families by identifying barriers to employment or reemployment prior to separation from active duty to include networking with One Stop Career Centers. For more information see below link or call the Texas RLL Liaison at 1-210-582-1622.
We strongly advise applicants to consult with the County Veterans Service Office prior to submitting this application. We can assist the widow in completing the application and acquiring supporting documents. This may help to reduce delays in receiving benefits or preventing the denial of benefits. |
IL-15 and IL-15 receptor selectively regulate differentiation of common mucosal immune system-independent B-1 cells for IgA responses.
We show in this report a new regulatory role for IL-15 and IL-15R in the development of B-1 cells and their differentiation into IgA-producing cells. Mucosal IgA levels were found to be inhibited by anti-IL-15 mAb treatment in vivo, but enhanced by administration of rIL-15, while serum IgA levels remained unaffected. Mucosal B-1 cells preferentially proliferated in response to IL-15 in vitro. When mucosal B-1 and B-2 cells were separated into surface (s)IgM(+)sIgA(-) and sIgM(-)sIgA(+) fractions, IL-15R-specific mRNA was found to be predominant in both sIgM(+)sIgA(-) and sIgM(-)sIgA(+) B-1 cells at a much higher level than B-2 cells. Further, incubation of these different subsets of B-1 and B-2 cells with IL-15 resulted in greater enhancement of the corresponding receptor expression by B-1 subset when compared with B-2 fraction. Interestingly, de novo isolated sIgM(+)sIgA(-) B-1, but not sIgM(+)sIgA(-) B-2, cells were already class-switched cells because the germline Calpha transcript was detected and was then further enhanced by IL-15. IL-15 also supported differentiation of both sIgM(+)sIgA(-) and sIgM(-)sIgA(+) B-1 cells into IgA-producing cells. Taken together, these findings suggest that IL-15 is a critically important cytokine for the differentiation of both sIgM(+),IgA(-) and sIgM(-)sIgA(+) B-1 cells expressing IL-15R into IgA-producing cells in mucosal tissues. |
Q:
Python read file line by line and print for debug
I'm trying to read a file line by line. This file has alot of lines.
Each line is formatted like for example:
data:otherdata1
data:someotherdata2
But for debug purposes I'm using a file with a few lines, same format.
And I'm trying to print those data.
But I'm getting the error:
print data.split(':')
^
SyntaxError: invalid syntax
What is wrong? I'm really lost in this.
def process(data):
print data.split(':') #printing for debug purposes
return data
file = open("list.txt")
for line in file:
process(line)
and yes I added import fileinput
A:
You are using Python3 with Python2 syntax.
In Python3, print is a function, not a statement. So use
print(data.split(':'))
instead.
|
//
// SendQueryEventArgs.cs
//
// Authors:
// Alan McGovern <alan.mcgovern@gmail.com>
//
// Copyright (C) 2019 Alan McGovern
//
// Permission is hereby granted, free of charge, to any person obtaining
// a copy of this software and associated documentation files (the
// "Software"), to deal in the Software without restriction, including
// without limitation the rights to use, copy, modify, merge, publish,
// distribute, sublicense, and/or sell copies of the Software, and to
// permit persons to whom the Software is furnished to do so, subject to
// the following conditions:
//
// The above copyright notice and this permission notice shall be
// included in all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
// EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
// LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
// OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
// WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
//
using System.Net;
using MonoTorrent.Dht.Messages;
namespace MonoTorrent.Dht
{
struct SendQueryEventArgs
{
public IPEndPoint EndPoint { get; }
public ErrorMessage Error { get; }
public Node Node { get; }
public QueryMessage Query { get; }
public ResponseMessage Response { get; }
public bool TimedOut => Response == null && Error == null;
public SendQueryEventArgs (Node node, IPEndPoint endpoint, QueryMessage query)
{
EndPoint = endpoint;
Error = null;
Node = node;
Query = query;
Response = null;
}
public SendQueryEventArgs (Node node, IPEndPoint endpoint, QueryMessage query, ResponseMessage response)
{
EndPoint = endpoint;
Error = null;
Node = node;
Query = query;
Response = response;
}
public SendQueryEventArgs (Node node, IPEndPoint endpoint, QueryMessage query, ErrorMessage error)
{
EndPoint = endpoint;
Error = error;
Node = node;
Query = query;
Response = null;
}
}
}
|
Tamaki niime
tamaki niime Co., Ltd. (有限会社玉木新雌 Yugen-Gaisha tamaki niime) is a Japanese apparel manufacturer of original Banshu-ori fabrics, based in Nishiwaki-shi, Hyogo. Banshu-ori fabric is made in one continuous production process, starting with a yarn to dyeing to sewing process. Popular products include the "Only One" Shawl that Niime Tamaki wove with a belt-type power loom made in 1965.
As of March 2017, their products are sold not only in Japan but also at 200 stores in 15 countries. A shawl produced by the company was also selected as "The Wonder 500" by the Ministry of Economy, Trade and Industry.
Summary
tamaki niime specializes in apparel produced using Banshu-ori fabric. Banshu-ori is a traditional brand of fabric manufactured in the North Harima area of Hyogo. The methods used to weave and dye the fabric are based on a technology brought to Hyogo by ”Miyadaiku” Yasubei Hida from Kyoto around 1800. Weaving and dyeing Banshu-ori was a popular off-season side job of farmers in the Nishiwaki-shi around Kyoto. Traditional Banshu-ori must conform to certain methods and quality standards to be considered authentic.
In 2004 fashion designer Niime Tamaki (originally a from Katsuyama-shi, Fukui) began working with Banshu-ori after a chance meeting with a Banshu-ori craftsman in Tokyo. Tamaki launched her own brand "tamaki niime" in 2004, which produced Banshu-or using both new manufacturing methods and traditional ones. For three years she designed and sold apparel in Osaka, using fabric woven by craftsman in Nishiwaki. In 2006 she decided to expand and established the company "tamaki niime".
The company moved to a store in Nishiwaki-shi, where they continued to weave and dye Banshu-ori. Tamaki started to do weaving in addition to her design activities and learned how to operate the loom. She experimented with the fabric, and developed a soft cloth that was difficult to sew and had the consistency of cotton candy. She wrapped it around her neck and noticed it was very comfortable to wear. This fabric became the “Only One” shawl, which was the main product of the company at the time. It was marketed as “an original shawl loved by everyone regardless of gender, age, or nationality”. The “Only One” shawl is woven using a belt-type power loom made in 1965, which creates a unique pattern and feeling that cannot be replicated using more modern machines.
The company continues to weave this fabric and apparel from it. The factory uses a variety of equipment to support the designing, dyeing, weaving, and sewing processes including a belt-type power loom, a rapier loom, yarn dyeing machine, warper, and CAD/CAM. They also use an arrange winder to create a rainbow-like gradation of multi-colored threads, as well as a state-of-the-art whole garment made by SHIMI SEIKI MFG LTD with the ability to knit without sewing.
tamaki niime make and produce shawls, shirts, pants, children’s clothes, bags and denims. They sell their products wholesale to partner select shops, department stores, and exhibitions all over Japan. Their works are marketed and sold internationally in over 200 stores in 15 countries including the United States, the United Kingdom, Canada, Mexico, New Zealand , and Taiwan.
The vision of tamaki niime is to turn the town of Hie-cho, where they are located, into a cotton town. In 2014 they began cultivating organic cotton on borrowed, abandoned farmland. They also are attempting to boost net domestic production of cotton by purchasing cotton seeds from growers at harvest and distributing them to people so they can also grow cotton. As they look to expand their business, Tamaki does consider issues of mass production, mass consumption, and global environmental impacts. In addition to cultivating cotton they grow vegetables which are used in their café annexed to their factory.
History
2004 – In December, Niime Tamaki makes brand “tamaki niime” aiming for new interpretation and development of Banshu-ori.
2006 – In April, “tamaki niime co., Ltd” established.
2008 – In April, opened directly managed store in Nishiwaki-shi.
2009 – In May, Tamaki moved to Nishiwaki-shi, and started development, announcement and production of original shawls.
2010 - In April, directly managed store “tamaki niime weaving room & stock room " opened in Nishiwaki-shi. In October, introduced two belt type weaving looms made in 1965. Started making "only one shawl" woven by Tamaki herself.
2011 - In May, introduced a Rapier loom made in 1983.
2012 – In March, introduced an innovative knitting machine.
2014 - In February, started cultivating organic cotton. In August, introduced the Netherlands hand weaving machine "louët" megado ", introduced a circular knitting machine in November.
2015 - "Hanayaka Kansai Selection 2016" and "Hyogo Female Future / Hanada Award" received
2016 - In September, moved "tamaki niime weaving room & stock room" to the foot of Mt. Okanoyama called "Japan's Navel" in Nishiwaki-shi Hie-town. Opened the newly renovated building which renovated the old dye factory about 5 times the area of the previous Lab. Aiming at "the space where the maker and the customer connect", it is structured so that Lab can be seen from Shop through the glass. On October 3, the Kansai Economic Federation selected Tamaki's "Roots Shawl" as one of ten points of Kansai's special product "Hanayaka Kansai Selection" that they would like to recommend foreigners.
2017 – In February, introduced two power looms made by 1967 and a warper. On February 21, exhibited at "Banshu-Ori Messe! 2017" (Minato-ku, Tokyo). In April, started organic cultivation of rice and vegetables. In October, introduced a circular knitting machine and two glass spinning machines. A total of 13 machines are in operation.
2018 - In January, newly established about 50 seats of food and drink space on the second floor of at the store and the Lab, offering a body-friendly lunch including organic coffee and locally produced tea, and offering a vegetarian menu called "haragoshirae-kai" on weekend.
()
References
External links
tamaki niime Official Website
tamaki niime Official Facebook
tamaki niime Instagram
Category:Companies based in Hyōgo Prefecture
Category:Japanese companies established in 2006
Category:Clothing brands of Japan
Category:Clothing companies established in 2006
Category:Textile companies of Japan
Category:Clothing companies of Japan |
Fat Bikes and Privilege
Note: in this climate, as a white woman it is a scary thing to try and write about race. If you have comments, would you please send them personally and let’s avoid a social media blitz. The following is part of my own journey as I struggle to speak for justice and love those who might be listening.
I love to run trails. Even in below zero, wind and snow, I layer up and run down the street and into Hartley. Yesterday I raced home from church and got 77 minutes of almost total bliss. Eyelashes frosted. Snot dripping. Yak tracks keeping my feet from slipping. And dodging fat bikes.
When you’re a runner and the fat bikes come at you, you get out of the way. At least I do. They could get off their bike and walk around me in the deep snow… but they never do. Never.
This isn’t a judgment on bikers, but it was impossible not to think of privilege and Martin Luther King Jr as I ran yesterday. Stay with me.
Fat bikes aren’t cheap. Neither are the fancy outfits most bikers wear in the winter while they are racing around the trails. And even if the bikers come from modest means they still never get off their bikes when I’m “sharing” the trail.
True Confession: At some point it is hard to trust the bikers. To say hello. To even like them while I am driving on roads and they are doing nothing to impede my day. And to not wonder if I were a man if it would be any different. Or as a woman I feel like this a lot… but I digress.
Perhaps if one or two of them did treat me differently it would change my attitude.
So as I ran yesterday, amazed by the tiny flakes of snow completely unique from the other billion falling around me. As I struggled to breathe as the weight of 4 layers of clothing kept me from my PR. As I wondered at the pure white snow… and how we love to think white is good and black is dark and bad. And as I dodged fat bikes, I asked for help.
Help in parenting my kids to treat everyone well. Help in continuing my education about how it is to be a minority in Duluth. Help in reading books like Waking up White, biographies written my Muslim women raising children in post 9-11 NYC and Tony Dungy’s experiences in the NFL in the ’70’s and ’80’s as a black coach. Help.
Help us Lord. It’s been 50 years since Martin Luther King Jr was killed. 50 years.
The next generation is watching. I know I was as I grew up in Dawson MN, pop 1626.
This is a picture of my father (center) shaking hands with Martin Luther King Sr at a Chapel meeting in 1972.
My father went to Duke Seminary when it was segregated. He tells stories of what it was like to try to get Duke to open their doors to African Americans. Inviting black pastors to speak at their chapel (which was “allowed”) and extending friendship.
I’ve always wanted a copy of this photo. My dad was an Army chaplain. He believed (still does) that Jesus is our model for how to treat others. And to stand up for what’s right. I’m pretty sure Martin Luther King Jr had a dad like that too. It’s what motivated him and created a movement that changed our country and the world.
I don’t know what the answers are all and I sit in a warm house today educated and well fed. But I do think how we treat one another, the words we use and our parenting will impact the world.
So as the state bleeds purple today, may we not forgot what day it is. |
Q:
Congratulations Screen on Android
I have a quiz app . when a user scores 90% and above the congratulation screen will be displayed. My question is i want the background of the screen to have small and different coloured paper like objects falling creating a illusion that i am congratulating him.
Please help me how to implement this idea.
A:
You can use animation to to show falling color objects. You can follow this link to view of this type of animation:-
Frame By Frame Animation
|
'#--
'# Copyright (c) 2006-2007 Luis Lavena, Multimedia systems
'#
'# This source code is released under the MIT License.
'# See MIT-LICENSE file for details
'#++
'##################################################################
'#
'# DO NOT INCLUDE THIS FILE DIRECTLY!
'# it is used internaly by ServiceFB
'# use ServiceFB.bi instead
'#
'##################################################################
namespace fb
namespace svc
'# now due references locking, I needed a constructor and destructor for
'# the namespace to garantee everything is cleaned up on termination of the process
declare sub _initialize() constructor
declare sub _terminate() destructor
'# global service procedures (private)
declare sub _main(byval as DWORD, byval as LPSTR ptr)
declare function _control_ex(byval as DWORD, byval as DWORD, byval as LPVOID, byval as LPVOID) as DWORD
declare sub _run()
'# global references helper
declare function _add_to_references(byref as ServiceProcess) as integer
declare function _find_in_references(byref as string) as ServiceProcess ptr
'# command line builder (helper)
'# this is used to gather information about:
'# mode (if present)
'# valid service name (after lookup in the table)
'# command line to be passed to service
declare sub _build_commandline(byref as string, byref as string, byref as string)
'# I started this as simple, unique service served from one process
'# but the idea of share the same process space (and reduce resources use) was good.
'# to do that, I needed a references table (similar to service_table, but we will
'# hold the ServiceProcess registered by ServiceHost (the multi services host).
'# also, I needed a locking mechanism to avoid problems of two calls changing the table
'# at the same time.
extern _svc_references as ServiceProcess ptr ptr
extern _svc_references_count as integer
extern _svc_references_lock as any ptr
end namespace '# fb.svc
end namespace '# fb
|
Observations and provocations from The Times' Opinion staff
McCain: Bomb, bomb Iran.... Oh, and Syria
March 7, 2012 | 1:48
pm
I've never been a big fan of those alternative-history novels in which Hitler wins World War II or Richard Nixon becomes president for life, but recent events have me pondering a hideous prospect: What if John McCain had defeated Barack Obama in 2008? The answer, as indicated by McCain's recent posturing, is that we'd be struggling with a lot more than an economic downturn; we'd probably be in costly and unwinnable wars not just in Afghanistan but in Syria and Iran.
McCain has not only forgotten the lessons of his own generation's war in Vietnam, he's forgotten what this generation learned in Iraq. He is eager not just for Israel to bomb Iran, which would set off a devastating regional conflict likely to drag in the United States, but for Washington to bomb Syria. On Monday, he became the first U.S. senator to call for air strikes on that country, and during a Senate Armed Services Committee meeting Wednesday, he admonished Defense Secretary Leon Panetta for failing to show leadership by "focusing on diplomatic and political approaches rather than a military intervention."
Panetta didn't take this sitting down; he said the administration was working to build international consensus, as it did in Libya, rather than taking unilateral action, and that as Defense secretary he has to know "what the mission is. I've got to make very sure we know whether we can achieve that mission, what price and whether or not it will make matters better or worse."
That's the part McCain either doesn't understand or doesn't care to discuss. U.S. military intervention in Syria in any form -- whether airstrikes or arming rebels -- would be extraordinarily risky. Syria is a powder keg of ethnic and sectarian factions with networks in neighboring countries; foreign intervention there would set off a proxy war that would further destabilize the entire Middle East.
To name just a few of the complications: In Lebanon, the politically powerful and heavily armed Hezbollah is committed to upholding the regime of Syrian President Bashar Assad, and it's not unrealistic to think that a broader civil war in Syria could spread to its fragile neighbor. If Assad should fall, it would almost certainly lead to reprisals, and likely atrocities, against Syria's minority Alawite community, the regime's most important domestic backers. The Syrian opposition that U.S. hawks would like to arm is an unknown quantity made up of Islamic fundamentalists and other groups that aren't necessarily sympathetic to U.S. interests. Taking out Syria's air defenses would be nowhere near as simple as taking out Libya's and would require a massive U.S. military commitment; it also presents risks that it would prompt Assad to use his country's stockpile of chemical weapons, which is said to be 100 times the size of Libya's.
I could go on, but I doubt I could say it better than the International Crisis Group, which wrote in a recent report:
Frustrated and lacking a viable political option, Western officials and analysts have toyed with a series of often half-baked ideas, from initiating direct military attacks to establishing safe havens, humanitarian corridors or so-called no-kill zones. All these would require some form of outside military intervention by regime foes that would more than likely intensify involvement by its allies. Even if they were to provoke the regime's collapse, that in itself would do nothing to resolve the manifold problems bequeathed by the conflict: security services and their civilian proxies increasingly gone rogue; deepening communal tensions; and a highly fragmented opposition.
McCain's hawkishness is starting to turn off most of his fellow Republicans, and even if he had won the White House, he might not have been able to fulfill his neocon nation-building fantasies. Fortunately, it will take an alternative-fiction writer, rather than a journalist, to imagine the harm he could have done. |
Q:
Prove that function $f$ is continuous at $x = x_{0}$
In class we're given the following definition about continuity, and I want to apply this definition to the problems that follow:
$f$ is continuous at $x_{0} \in \mathrm{dom}(f)$ if $\forall x_{n} \in \mathrm{dom}(f)$ that's converging to $x_{0}$, $\lim_{n \rightarrow \infty} f(x_{n}) = f(x_{0})$
$1). $ Prove that $f(x) = \sqrt{x}$ is continuous at $x_{0} = 0$.
Suppose $\lim_{n \rightarrow \infty} (x_{n}) = 0$, then
$\lim_{n \rightarrow \infty} f(x_{n}) = \lim_{n \rightarrow \infty} \sqrt{x_{n}} = (\lim_{n \rightarrow \infty} x_{n})^{\frac{1}{2}} = 0^{\frac{1}{2}} = f(x_{0})$
And note that $\lim_{n \rightarrow \infty} \sqrt{x_{n}} = (\lim_{n \rightarrow \infty} x_{n})^{\frac{1}{2}}$ holds because $f$ is a continuous function. From my understanding, I can apply the algebraic limit theorem only if the function is continuous, correct? But here I'm using it as part of my proof to show that the function is continuous at $0$. Is this a problem? Or can the algebraic limit theorem be applied regardless of continuity?
$2).$ Prove that $f(x) = \begin{cases}
x \sin(\frac{1}{x})) & \text{ if } x \neq 0 \\
0 & \text{ otherwise }
\end{cases}$
Suppose $\lim_{n \rightarrow \infty} x_{n} = 0$, then
$\lim_{n \rightarrow \infty} f(x_{n}) = \lim_{n \rightarrow \infty} 0 = 0 = f(x_{0})$
Would this suffice? Should I also take into account that $x_{n}$ might consist of $x \neq 0$?
A:
For your first question: no, because the property $\lim \sqrt{x_n}=\sqrt{\lim x_n}$ can be proved by definition of limit of sequence. Indeed, if $x_n\ge 0$ for all $n$ and $\lim x_n=L\ge 0$ then $\lim \sqrt{x_n}=\sqrt{L}$. If $L=0$, for all $\varepsilon>0$ there exists $N$ such that $x_n<\varepsilon^2$ for all $n>N$, then $\sqrt{x_n}<\varepsilon$ for all $n>N$, thus $\lim \sqrt{x_n}=0$. If $L>0$, we have $|\sqrt{x_n}-\sqrt{L}|=\frac{|x_n-L|}{\sqrt{x_n}+\sqrt{L}}<\frac{|x_n-L|}{\sqrt{L}}$, so $\lim \sqrt x_n=\sqrt{L}$.
For your second one, yes, $x_n\neq 0$ for all $n$, since this is the definition of continuity of a function: $f$ is continuous at $x_0$ if and only if for all sequence $(x_n)$ such that $x_n\neq x_0$, $\lim f(x_n)=f(x_0)$.
Due to my experience, it's often easier to use the above definition to prove the continuity of function than the $\varepsilon-\delta$ one.
|
Unfortunately, young and inexperienced drivers are known to be at higher risk of having an accident and making a claim on their insurance.
The road safety charity Brake has research showing that drivers aged between 17 and 19 years old make up just 1.5% of all UK drivers, but have around 9% of fatal or serious road accidents when they are driving. 16 to 19 year-olds are a third more likely to die in a crash than older drivers, and one in four 18 to 24 year-olds crash within two years of passing their driving test.
Insurers base insurance premiums on risk, so it's not surprising that this kind of data can increase the cost of insurance for young drivers under the age of 25, and perhaps more so for new drivers under 21. And if you've been unfortunate enough to have had an accident or a conviction, you might find it even harder to get on the road.
We can help you find a cheaper premium. No fee, no hassle, just affordable car insurance when you need it most. |
Resolving the problem
This defect will not impact the main usage of Flume, which is to transfer data. You can also get node configuration from Flume master console. Because the Flume latest version (1.2.0) architecture has changed, there is no Flume master any more. This issue does not exist in Flume NG. This is a fixed upstream issue. |
Enjoy the best and freshest of Phuket Seafood at Momo Cafe with our special promotion of 4 dine for the price of 3 at our Friday Night Buffet.
Treat yourself to a diverse selection of the freshest local seafood delicacies including Phuket Tiger prawn and Andaman Blue crab as well as a diverse assortment of high quality imported lamb, beef and mussels from New Zealand and Australia.
Price per person is set at BHT 650++ and is inclusive of food only, children aged 6-12 pay half price while kids under 5 eat for free. |
1. Introduction {#s1}
===============
For a long time, the brain-computer interface (BCI) has been a prevalent communication system for directly bridging between a brain and a computer. A typical BCI system collects brain activities associated with mental tasks, translates these neural signals into appropriate commands, and eventually sends them to a computer (Handiru and Prasad, [@B10]). As a non-invasive brain activity measurement method, electroencephalography (EEG) has attracted increasing interest, owing to its low risk, low cost, feasibility, and significant potential for practical applications (Yang et al., [@B39]). EEG data is commonly composed of multichannel signals recorded from several electrodes placed on the scalp to obtain the activity of various cortexes. For example, motor-imagery (MI) EEG signals are primarily gained from the motor-relevant cortex and could provide users with direct control of various devices (Lafleur et al., [@B18]; Meng et al., [@B25]), e.g., a wheelchair, quadcopter, or robotic arm, without using any peripheral nerves or muscle movements.
The core component of an EEG-based BCI (EEG-BCI) system is the decoding or classification of EEG signals. The basic structure of an EEG classification framework typically consists of four parts: signal pre-processing, feature extraction, channel selection and classification, of which the feature extraction and channel selection are attracting more attention. Feature extraction is fundamental to EEG classification and many methods have been presented to acquire the implicit essential information from raw EEG signals, e.g., spectral power (SP) and time-domain parameters (TDP) (Müller et al., [@B30]). Channel selection attempts to remove irrelevant EEG features in redundant channels to reduce the setup time and equipment cost, and improve the effectiveness and efficiency of EEG-BCI systems (Alotaiby et al., [@B1]). Well-known channel selection methods are often based on evolutionary algorithms and mutual information.
Even though multiple methods have been proposed to tackle EEG classification tasks using different feature extraction and channel selection methods, bottlenecks still appear when existing EEG-BCI frameworks are applied to practical applications, as shown in Figure [1A](#F1){ref-type="fig"}. (1) The computational complexity of channel selection methods has often been neglected by previous works (Kee et al., [@B16]; Zhang et al., [@B42]), thereby slowing the EEG feature extraction and offline classifier training procedures, which negatively affects the deployment efficiency of EEG frameworks in different situations. Moreover, excessively long setup and experiment times may negatively affect the (human) subjects\' mental concentration and condition, heavily decreasing the EEG signal-to-noise ratio (SNR). (2) Convergence is seldom considered in some iterative frameworks (Yu et al., [@B40]), leading to stochastic, even divergent solution searching. A non-monotonic convergence process may postpone the EEG classification and further degenerate the subjects\' experience. Moreover, it is difficult to consistently obtain an optimal decision persistently because of the unstable solution searching process. (3) Existing frameworks are mostly designed too specifically (Bashar et al., [@B2]), narrowing their expansibility. The majority of datasets in pattern recognition approaches are expressed in two-dimensional matrices, i.e., samples and features, which differ from previous EEG data matrices. Thus, an overly well-designed EEG framework is usually confined to specific methods and offers few open accesses to the many more remaining methods, thus, seriously affecting its application and expansion. Additionally, considering that novel methods are continuously emerging, inadequate support for them may limit the sustained improvement of existing frameworks.
{#F1}
Intuitively, the solution to (2) is to use fast convergent channel selection algorithms, which is also beneficial for solving (1). Another solution to (1) is decreasing the data scale, including feature extraction and dimension reduction. Meanwhile, the solution to (3) is to present the EEG signals in two dimensional data matrices, which are commonly used in other signal processing areas, without removing the latent spatial information carried by multi-channel EEG signals. In the meantime, representing EEG data in two dimensional matrices is conducive to building bridges between EEG feature extraction and previous feature selection methods.
In light of this, as shown in Figure [1A](#F1){ref-type="fig"}, we propose a fast, open EEG classification framework characterized by feature compression, low-dimensional representation, and convergent iterative channel ranking. Firstly, to alleviate the computational burden, we use fast data clustering to give EEG feature vectors numerical signatures according to their channel-wise similarity. Thus, it is possible to reduce the feature dimension in EEG signals. Secondly, instead of simply flattening three dimensional (trial-channel-feature) EEG signals into two dimensional (trial-channel^\*^feature) ones, we compress the EEG features and contract the feature vectors into their numerical signatures. Thus, we can represent the EEG signal using a two dimensional matrix and retain its spatial information in the meantime. In this manner, EEG signals are mapped into a normal data matrix with rows and columns indicating trials and channels, respectively, which is consistent with most pattern recognition studies. Thirdly, to obtain fast and stable decisions, we use a few iterative feature selection approaches with theoretical convergence to rank and select informative EEG channels, thereby shortening the procedure and improving the decision performance of the EEG-BCI systems.
Our main aims and contributions are highlighted in the following: We leverage the data clustering between feature extraction and channel selection in the traditional EEG-BCI to channel-wise compress the features to reduce the data scale.We provide an EEG classification framework applicable to most common EEG processing and pattern recognition methods to provide an entrance for increasing the EEG-BCI performance through technology integration.We introduce feature selection approaches with theoretical convergence to rank and select channels to further improve the performance of the EEG-BCI.
2. Related work {#s2}
===============
In past decades, great efforts have been made to address the EEG classification problem, mainly including feature extraction and channel selection.
One of the most widely used EEG feature extraction methods is common spatial patterns (CSP). CSP utilizes covariance analysis to amplify the class disparity in the spatial domain, to combine signals from different channels (Blankertz et al., [@B5]; Li et al., [@B21]). Improved CSPs, e.g., common spatio-spectral patterns (CSSP) (Lemm et al., [@B20]), iterative spatio-spectral pattern learning (ISSPL) (Wu et al., [@B36]), and filter bank common spatial patterns (FBCSP) (Kai et al., [@B15]), mainly optimize the combination of multi-channel signals by developing a spectral weight coefficient evaluation. Obviously, these spatial feature extraction methods are good at selecting the pivotal spatial information included in EEG signals. However, signal combination cannot be practically used to reduce the channel number and data scale. Apart from the spatial domain, extracting features in the temporal and frequency domains is also prevalent in many works. Multivariate empirical mode decomposition (MEMD) generates multiple dimensional envelopes by projecting signals in all directions of various spaces (Rehman and Mandic, [@B31]; Islam et al., [@B13]). The autoregressive (AR) model assumes that EEG signals can be approximated in the AR process, in which the features could be gained as parameters of the approximated AR models (Zabidi et al., [@B41]). TDP was introduced as a set of broadband features based on the variance of different EEG signals in various orders (Vidaurre et al., [@B34]). A wavelet transform (WT) simultaneously provides ample frequency and time information about EEG signals at the low and high frequencies, respectively (Jahankhani et al., [@B14]). We utilize SP based on AR and TDP as two feature extraction examples in this paper.
EEG channel selection has been extensively studied. For instance, multi-objective genetic algorithms (GA) (Kee et al., [@B16]) and Rayleigh coefficient maximization based GA (He et al., [@B11]) were introduced to simultaneously optimize the number of selected channels and improve the system accuracy by embedding classifiers into the GA process. Recursive feature elimination (Guyon and Elisseeff, [@B7]) and zero-norm optimization (Weston et al., [@B35]) based on the training of support vector machines (SVMs) were used to reduce the number of channels without decreasing the motor imagery EEG classification accuracy (Lal et al., [@B19]). Sequential floating forward selection (SFFS) (Pudil et al., [@B28]) and successive improved SFFS (ISFFS) (Zhaoyang et al., [@B43]) took an iterative channel selection strategy that selected the most significant feature from the remaining features and dynamically deleted the least meaningful feature from the selected feature subset. A Gaussian conjugate group-sparse prior was incorporated into the classical empirical Bayesian linear model to gain a group-sparse Bayesian linear discriminant analysis (gsBLDA) method for simultaneous channel selection and EEG classification (Yu et al., [@B40]). Mean ReliefF channel selection (MRCS) adopted an iterative strategy to adjust the ReliefF-based weights of channels according to their contribution to the SVM classification accuracy (Zhang et al., [@B42]). The Fisher criterion, based on Fisher\'s discriminant analysis was utilized to evaluate the discrimination of TDP features, extracted from all channels in different time segments via channel selection, using time information (CSTI) methods (Yang et al., [@B39]). GA and pattern classification using multi-layer perceptrons (MLP) and rule-extraction based on mathematical programming were combined to create a generic neural mathematical method (GNMM) to select EEG channels (Yang et al., [@B38]). However, a visible limitation is that the convergence of most of these methods is unstable, e.g., gsBLDA, leading to an uncertain channel selection procedure and selected subset.
Therefore, it is natural to select EEG channels by taking advantage of feature selection methods in other areas, e.g., robust feature selection (RFS) (Nie et al., [@B26]), joint embedding learning and sparse regression (JELSR) (Hou et al., [@B12]), nonnegative discriminative feature selection (NDFS) (Li et al., [@B22]), selecting feature subset with sparsity and low redundancy (FSLR) (Han et al., [@B9]), robust unsupervised feature selection (RUFS) (Qian and Zhai, [@B29]), joint Laplacian feature weights learning (JLFWL) (Yan and Yang, [@B37]), a general augmented Lagrangian multiplier (FS_ALM) (Cai et al., [@B6]) and structural sparse least square regression based on the *l*~0~-norm (SSLSR) (Han et al., [@B8]). RFS, JELSR, and NDFS focused on the *l*~2,\ 1~-norm minimization regularization to develop an accurate and compact representation of the original data. RUFS tried to solve the combined object of robust clustering and robust feature selection using a limited-memory BFGS based iterative solution. JLFWL selected important features based on *l*~2~-norm regularization, and determined the optimal size of the feature subset according to the number of positive feature weights. FSLR could retain the preserving power, while implementing high sparsity and low redundancy in a unified manner. FS_ALM and SSLSR attempted to handle the least square regression based on *l*~0~-norm regularization by introducing a Lagrange multiplier and a direct greedy algorithm, respectively. Further, prevalent deep learning was also utilized in this area. A point-wise gated convolutional deep network was developed to dynamically select key features using a gating mechanism (Zhong et al., [@B44]). Multi-modal deep Boltzmann machines were employed to select important genes (biomarkers) in gene expression data (Syafiandini et al., [@B33]). A deep sparse multi-task architecture was exploited to recursively discard uninformative features for Alzheimer\'s disease diagnosis (Suk et al., [@B32]). In this paper, we employ three performance-verified feature selection methods with theoretical convergence, i.e., RFS, RUFS, and SSLSR, to rank and select channels.
3. Materials and methods {#s3}
========================
In this section, we describe the datasets, introduce some notations used throughout this paper, and present the proposed frameworks.
3.1. Datasets and pre-processing
--------------------------------
In our experiments, we selected two public real world datasets on motor imagery (MI) paradigms as examples. A brief description of these two datasets is offered below and their basic statistics are summarized in Table [1](#T1){ref-type="table"}. EEG datasets of other paradigms, e.g., P300, are also suitable for this framework.
######
Statistics of the two datasets.
**Datasets** **DS1** **DS2**
---------------------------- ---------------- ---------
Number of channels 118 59
Sampled frequency (Hz) 100 100
Number of subjects 5 4
Subjects\' name aa,al,av,aw,ay a,b,f,g
Number of trials per class 140 100
### 3.1.1. DS1
Dataset IVa from BCI Competition III is a public EEG dataset provided by the Berlin BCI group Fraunhofer FIRST (Intelligent Data Analysis Group) and Campus Benjamin Franklin of the *Charité* University (Neurophysics Group). This public dataset is recorded from five healthy subjects (aa,al,av,aw,ay) during right hand and right foot motor imageries. The EEG recordings consist of 118 channels at positions of the extended international 10/20-system. We choose a version of the data that is down-sampled at 100 Hz for analysis. In the experiments, subjects performed three motor imageries for 3.5 s after visual cues for left hand, right hand, or right foot. After the duration of motor imagery, a resting interval with random length of 1.75--2.25 s was inserted for relaxation. The dataset provides only EEG trials for right hand and right foot imagery. For each subject, the dataset consists of signals of 140 trials per class.
In our experiment, the signals in the time interval of \[0.5, 3.0\] s are analyzed for each trial. A visual cue is presented to mark the start time (0 s). A bandpass finite impulse response (FIR) filter using the window method, with a band of 0.1--40 Hz and order 33, is applied to DS1.
### 3.1.2. DS2
Dataset I from BCI Competition IV is another public EEG dataset provided by the Berlin BCI group Fraunhofer FIRST (Intelligent Data Analysis Group) and Campus Benjamin Franklin of the *Charité* University (Neurophysics Group). This public dataset is recorded from four healthy subjects (a,b,f,g) during two classes of motor imagery selected from three classes: left hand, right hand, and foot (side chosen by the subject; optionally also both feet). In the experiments, the data was continuous signals of 59 EEG channels and visual cues pointing left, right or down were presented for a period of 4.0 s during which the subject was instructed to perform the cued motor imagery task. These periods were interleaved with 2.0 s of blank screen and 2.0 s with a fixation cross displayed in the center of the screen. The dataset provides only EEG trials for left hand and foot imagery. For each subject, the dataset consists of signals of 100 trials per class.
In our experiment, the signals in the time interval of \[0.0, 4.0\] s are analyzed for each trial. A visual cue is presented to mark the start time (0 s). A bandpass FIR filter using the window method, with a band of 0.1 to 40 Hz and order 33, is also applied to DS2.
3.2. Notations
--------------
In this document, scalars, matrices, vectors, sets, and functions are denoted as small, boldface capital, boldface lowercase, blackboard capital, and script capital letters, respectively. **x**^*T*^, **X**^*T*^, **x**~*i*~, **X**~*i*~, **X**~*ij*~, **X**~(*i*,:)~, **X**~(:,\ *j*)~, and trace(**X**)indicate the transpose of vector **x**, the transpose of matrix **X**, the *i*-th element of **x**, the *i*-th sample of the variable **X**, the element of **X** occurring in the *i*-th row and *j*-th column, the *i*-th row of **X**, the *j*-th column of **X** and the trace of **X** respectively. Moreover, \|\|**x**\|\|~1~ is the *l*~1~-norm of **x**, \|\|**X**\|\|~1~ and \|\|**X**\|\|~2~ are the *m*~1~-norm and *m*~2~-norm of matrix **X**. For any vector **x** ∈ ℝ^*n*×1^ and any matrix **X** ∈ ℝ^*n*×*m*^, the definitions of *l*~*p*~-norm, *m*~*r*~ -norm and *m*~*r,\ s*~ -norm are given in the Appendix.
Furthermore, assume that we have recorded EEG signals of *N* trials, and let $\mathbb{X} = \left\{ \mathbf{\text{X}}_{i} \right\}_{i = 1}^{N}$ be the set of EEG signal corresponding to all *N* trials. Specifically, we represent each trial of EEG signals, i.e., **X**~*i*~(*i* = 1, 2, ..., *N*), as matrix **X**^*M*×*C*^, where *M* and *C* are the number of sampled time points and channels in a trial respectively. The class indicator vector can be denoted as **y** ∈ {0, 1, ..., *L* − 1}^*N*^, where *L* is the number of classes.
3.3. Proposed methods
---------------------
We first depict the flowchart of the framework, called feature compressing and channel ranking (FCCR). Afterward, we provide the details for its three core components. Lastly, its pseudo codes are given.
### 3.3.1. Flowchart of the framework
Our framework starts with EEG signal pre-processing and single-channel SP or TDP feature extraction. Next, after dividing all of the trials into the training and testing sets, we gather all the features from the different channels for the training trials and assign them cluster signatures through *k*-means. Then, EEG signals are mapped from the three dimensional (trial-channel-feature) to the two dimensional (trial-channel) matrix by compressing all feature vectors into their cluster signatures. With this data representation (two-dimensional matrix) commonly used in most pattern recognition areas, we employ RFS, RUFS, or SSLSR to rank and select the channels. Lastly, a traditional LDA is used to classify the testing trials, with the features corresponding to the selected channels. The procedure of our proposed FCCR is summarized in Figure [1B](#F1){ref-type="fig"}. Notably, FCCRs are subject-specific since the procedures from feature clustering to channel ranking in Figure [1B](#F1){ref-type="fig"} are dependent of the subject.
### 3.3.2. Feature extraction
We take SP based on AR model and TDP as two examples to extract EEG features. In fact, other methods, including WT and MEMD, can also be used to replace them.
Each EEG signal of a single channel is treated as a time-varying variable, denoted as **x**(*t*), in this section.
#### 3.3.2.1. SP features
A power spectrum is an energy density distribution over frequencies. The AR Filter actually computes a time series of such spectra, representing a flow of energy through each single point in the frequency domain. In this way, the SP based on AR model can be treated as the energy per frequency per time.
AR model is a representation of a type of random process (Zabidi et al., [@B41]), which specifies that the output variable depends linearly on its own previous values. Hence, the predicted value $\hat{\mathbf{\text{x}}}{(t)}$ of an AR process is defined by,
x
\^
(
t
)
=
\-
∑
k
=
1
p
a
p
(
k
)
x
(
t
\-
k
)
,
where *a*~*p*~(*k*) is the coefficient of the *p*-th order AR process.
Then, the estimated power spectrum directly corresponds to the filter\'s transfer function whose coefficients are actually AR coefficients. To obtain spectral power for finite-sized frequency bins, that power spectrum is multiplied by total signal power, and integrated over the frequency ranges corresponding to individual bins. We select frequency bins from 0 to 30 Hz with step 3 Hz, resulting 11 values per channel for each trial.
#### 3.3.2.2. TDP features
TDPs are a set of broadband features with physical meanings, whose definition is as (Vidaurre et al., [@B34]).
T
D
P
(
p
)
=
log
(
var
(
d
p
x
(
t
)
d
t
p
)
)
,
p
=
0
,
1
,
2
,
...
,
in which we apply logarithm to regulate the TDPs\' distribution to Gaussian approximately (Vidaurre et al., [@B34]) to fit the LDA classifier (Müller et al., [@B30]).
Three TDPs are used in this paper to extract EEG features in both temporal and frequency domain, that is, **f**^*TDP*^ = \[*TDP*^(0)^, *TDP*^(1)^, *TDP*^(2)^\]. Typically, *TDP*^(0)^ depicts the EEG character in terms of amplitude, *TDP*^(1)^ can be interpreted as a kind of EEG pattern in terms of high frequency (mainly the beta band), and *TDP*^(2)^ carries the essential information of the change in frequency (Vidaurre et al., [@B34]).
### 3.3.3. Feature compressing and data representation
With SP or TDP features extracted from all channels of all trials, we collect all features of training trials (with $\overline{N}$ trials) as a feature pool ($\left\{ \mathbf{\text{f}}_{i} \right\}_{i = 1}^{\overline{N} \times C \times 11}$ for SP features or $\left\{ \mathbf{\text{f}}_{i} \right\}_{i = 1}^{\overline{N} \times C \times 3}$ for TDP features), abandoning its trial and channel information. Then, the classical *k*-means is used as an example to cluster the above EEG features. Likewise, almost all of data clustering approaches, such as spectral clustering and AP clustering, are acceptable to replace it.
In this paper, we utilize *k*-means clustering to partition $\overline{N} \times C \times 11$ or $\overline{N} \times C \times 3$ features into $K{({\leq \overline{N}})}$ sets ℂ = {ℂ~1~, ℂ~2~, ..., ℂ~*K*~} in order to minimize the within-cluster sum of squares (WCSS). Formally, the object of *k*-means is to find:
argmin
ℂ
∑
j
=
1
K
∑
i
∈
ℂ
j
\|
\|
f
i
\-
μ
j
\|
\|
2
,
where μ~*j*~ is the mean of features in ℂ~*j*~. This is equivalent to minimizing the pairwise squared deviations of features in the same cluster:
argmin
ℂ
∑
j
=
1
K
1
2
\|
ℂ
j
\|
∑
i
,
i
′
∈
ℂ
j
\|
\|
f
i
\-
f
i
′
\|
\|
2
After dividing training features into *K* clusters using standard Lloyd\'s algorithm (Lloyd, [@B23]), we can replace feature vector **f**~*i*~ with its corresponding cluster signature *j* if *i*∈ℂ~*j*~. In this way, training signals can be represented as a matrix ${\overset{\sim}{\mathbf{\text{X}}}}^{\overline{N} \times C}$.
### 3.3.4. Channel ranking
After representing training trials as a normal data matrix $\overset{\sim}{\mathbf{\text{X}}}$ with its rows and columns indicating trials and channels, we take RFS, RUFS and SSLSR as examples to rank and select EEG channels. The main strength of these methods is that they could find the optimal channel subset with the strict convergence in theory. Similarly, researchers can use other appropriate feature selection algorithms to rank channels.
In this section, $\overset{\sim}{\mathbf{\text{Y}}} \in \left\{ {0,1} \right\}^{\overline{N} \times L}$ is the class label indicator matrix (${\overset{\sim}{\mathbf{\text{Y}}}}_{il} = 1$ if the *i*-th sample belongs to the *l*-th class, otherwise, ${\overset{\sim}{\mathbf{\text{Y}}}}_{il} = 0$), and **W** ∈ ℝ^*C*×*L*^ is the weight matrix of channels. After obtaining **W**, we rank all channels according to their weights, which are computed as \|\|**W**~(*i*,:)~\|\|~2~.
#### 3.3.4.1. Ranking based on RFS
RFS is an efficient and robust feature selection by emphasizing *l*~2,\ 1~-norm minimization on both the loss function and the regularization term (Nie et al., [@B26]). The objective of RFS is
min
W
J
(
W
)
=
\|
\|
X
\~
W
\-
Y
\~
\|
\|
2
,
1
\+
α
\|
\|
W
\|
\|
2
,
1
.
Note that Equation (5) is equivalent to
min
W
,
E
\|
\|
E
\|
\|
2
,
1
\+
\|
\|
W
\|
\|
2
,
1
,
s
.
t
.
X
\~
W
\+
α
E
=
Y
\~
.
By setting $\mathbf{\text{A}} = \left\lbrack {\overset{\sim}{\mathbf{\text{X}}}\ \alpha\mathbf{\text{I}}} \right\rbrack \in {\mathbb{R}}^{\overline{N} \times {({C + \overline{N}})}}$ and $\mathbf{\text{U}} = \left\lbrack \left. \begin{matrix}
\mathbf{\text{W}} \\
\mathbf{\text{E}} \\
\end{matrix} \right\rbrack \right. \in {\mathbb{R}}^{{({C + \overline{N}})} \times L}$, we can reformulate the objective of RFS as
min
U
\|
\|
U
\|
\|
2
,
1
\+
\|
\|
W
\|
\|
2
,
1
,
s
.
t
.
A
U
=
Y
\~
.
Then, the optimal **U** is gained by setting the derivative of the Lagrangian function of Equation (7) on **U** to zeros, that is
U
=
D
\-
1
A
T
(
A
D
\-
1
A
T
)
\-
1
Y
\~
,
where **D** is a diagonal matrix with the *i*-th diagonal element as $\mathbf{\text{D}}_{ii} = \frac{1}{\left. 2 \middle| \middle| \mathbf{\text{U}}_{({i,:})} \middle| |_{2} \right.}$.
The pseudo codes of RFS is given in Algorithm 1.
######
Algorithm 1 Pseudo codes of RFS
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**Input:**
EEG data matrices ${\overset{\sim}{\mathbf{\text{X}}}}^{\overline{N} \times C}$, ${\overset{\sim}{\mathbf{\text{Y}}}}^{\overline{N} \times L}$ and hyper-parameter α.
**Output:**
Channel weight matrix **W**.
1: Set *iter* = 0, $\mathbf{\text{A}} = \left\lbrack {\overset{\sim}{\mathbf{\text{X}}}\ \alpha\mathbf{\text{I}}} \right\rbrack$ and initialize $\mathbf{\text{D}}_{0}^{{({C + \overline{N}})} \times {({C + \overline{N}})}}$ as an identity matrix.
2: **repeat**
3: Calculate $\mathbf{\text{U}}_{iter + 1} = \mathbf{\text{D}}_{iter}^{- 1}\mathbf{\text{A}}^{T}{({\mathbf{\text{A}}\mathbf{\text{D}}_{iter}^{- 1}\mathbf{\text{A}}^{T}})}^{- 1}\overset{\sim}{\mathbf{\text{Y}}}$.
4: Update **D**~*iter*+1~, where the *i*-th diagonal element as $\mathbf{\text{D}}_{ii} = \frac{1}{\left. 2 \middle| \middle| \mathbf{\text{U}}_{({i,:})} \middle| |_{2} \right.}$.
5: iter = iter + 1.
6: **until** *Converges*
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
The theoretical proof of the convergence of Algorithm 1 can be found in Nie et al. ([@B26]).
#### 3.3.4.2. Ranking based on RUFS
RUFS utilizes *l*~2,\ 1~-norm minimization on processes of both label learning and feature learning to effectively handle outliers and noise in the data, as well as reduce redundant or irrelevant channels. The objective of RUFS is often written as
min
F
,
G
,
W
\|
\|
X
\~
\-
G
F
\|
\|
2
,
1
\+
β
trace
(
G
T
L
G
)
\+
γ
\|
\|
X
\~
W
\-
G
\|
\|
2
,
1
\+
η
\|
\|
W
\|
\|
2
,
1
,
s.t.
G
∈
ℝ
\+
N
¯
×
L
,
G
T
G
=
I
,
F
∈
ℝ
\+
L
×
C
,
W
∈
ℝ
\+
C
×
L
,
where **L** is the normalized graph Laplacian matrix commonly used in unsupervised learning (Li et al., [@B22]).
Then, the optimal **W** is obtained by using a limited-memory BFGS (Nocedal, [@B27]) and BMLVM (Benson and Moré, [@B3]) based alternating iterative algorithm, which can be summarized in Algorithm 2.
######
Algorithm 2 Pseudo codes of RUFS
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**Input:**
EEG data matrix ${\overset{\sim}{\mathbf{\text{X}}}}^{\overline{N} \times C}$ and hyper-parameters β,γ,η.
**Output:**
Channel weight matrix **W**.
1: Set *iter* = 0 and initialize **G**~0~, $\left. \mathbf{\text{F}}_{0}\leftarrow\left\lbrack {{({\mathbf{\text{G}}^{T}\mathbf{\text{G}}})}^{- 1}\mathbf{\text{G}}^{T}\overset{\sim}{\mathbf{\text{X}}}} \right\rbrack_{+} \right.$, **W**~0~.
2: **repeat**
3: Fixing **G**~*iter*~, compute **W**~*iter*+1~ from L-BFGS algorithm given **G**~*iter*~, **W**~*iter*~, β and γ.
4: Fixing **F**~*iter*~ and **W**~*iter*+1~, compute **G**~*iter*+1~ from BMLVM algorithm given **G**~*iter*~, **F**~*iter*~, β and η.
5: Fixing **G**~*iter*+1~, compute **F**~*iter*+1~ from BMLVM algorithm given **G**~*iter*+1~, **F**~*iter*~.
6: iter = iter + 1.
7: **until** *Converges*
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
The theoretical proof of the convergence of Algorithm 2 is referenced in Qian and Zhai ([@B29]) and Bezdek and Hathaway ([@B4]).
#### 3.3.4.3. Ranking based on SSLSR
SSLSR is an effective greedy algorithm to directly handle the challenging *l*~2,\ 0~-norm based structural sparse least square regression, which has the objective
min
W
L
(
W
)
:
min
W
1
N
¯
\|
\|
X
\~
W
\-
Y
\~
\|
\|
2
2
,
s.t.
\|
\|
W
\|
\|
r
,
0
≤
C
¯
,
where $\overline{C}$ is the number of selected channels.
Start with **W**~0~ = **0**, SSLSR develops an alternating forward and backward strategy to pick and eliminate (if possible) a channel in each iteration.
Suppose at the beginning of the *iter*-th iteration, the selected channel set is 𝔽^*iter*^. Then $\forall i \in \mathbb{F}^{iter},\mathbf{\text{W}}_{({i,:})}^{iter} \neq \mathbf{0}$. In the forward step, we need to find the channel which reduces the loss most, i.e., find *i* ∉ 𝔽^*iter*^ and θ∈ℝ^*L*×1^ such that
min
i
∉
𝔽
i
t
e
r
min
θ
L
(
W
i
t
e
r
\+
e
i
θ
T
)
,
where $\mathbf{\text{e}}_{i} \in {\mathbb{R}}^{C \times 1}$ is the vector of zeros, except for the *i*-component which is one. θ is the optimal weight vector for one channel.
Since $L{({\mathbf{\text{W}}^{iter} + \mathbf{\text{e}}_{i}\theta^{T}})}$ is convex in terms of θ^*T*^, the optimal θ for *i* ∉ 𝔽^*iter*^ can be gained by setting $\partial L{({\mathbf{\text{W}}^{iter} + \mathbf{\text{e}}_{i}\theta^{T}})}/\partial\theta = 0$, that is,
θ
=
(
Y
\~
\-
X
\~
W
i
t
e
r
)
T
X
\~
(
:
,
i
)
/
\|
\|
X
\~
(
:
,
i
)
\|
\|
2
2
.
Thus, we have
\[
i
i
t
e
r
,
θ
i
t
e
r
\]
=
argmin
i
∉
𝔽
i
t
e
r
argmin
θ
L
(
W
i
t
e
r
\+
e
i
θ
T
)
.
Before beginning the backward step, 𝔽 and **W** are updated as 𝔽^*iter*+1^ = 𝔽^*iter*^∪{*i*^*iter*^} and $\mathbf{\text{W}}^{iter + 1} = \mathbf{\text{W}}^{iter} + \mathbf{\text{e}}_{i^{iter}}\theta^{iterT}$ respectively.
In the backward step, the channel with the least contribution to reducing the loss is selected and we need to determine whether to remove it according to a specific criterion. That is to say, we need firstly to find
j
i
t
e
r
=
argmin
j
∈
𝔽
i
t
e
r
L
(
W
i
t
e
r
\-
e
j
W
(
j
,
:
)
i
t
e
r
)
\-
L
(
W
i
t
e
r
)
,
which is equivalent to finding *j* ∈ 𝔽^*iter*^ such that
min
j
∈
𝔽
i
t
e
r
L
(
W
i
t
e
r
\-
e
j
W
(
j
,
:
)
i
t
e
r
)
.
The eliminating criterion relies on the decrease and increase of the loss in forward and backward steps. Specifically, if and only if
Δ
\+
i
t
e
r
\<
ν
Δ
\-
i
t
e
r
,
the *j*^*iter*^-th channel can be removed from 𝔽, where ν∈(0, 1) is a hyper-parameter, $\Delta_{-}^{iter} = L{(\mathbf{\text{W}}^{iter})} - L{(\mathbf{\text{W}}^{iter + 1})}$ and $\Delta_{+}^{iter} = L{({\mathbf{\text{W}}^{iter} - \mathbf{\text{e}}_{j^{iter}}\mathbf{\text{W}}_{({j^{iter},:})}^{iter}})} - L{(\mathbf{\text{W}}^{iter})}$.
The pseudo codes of SSLSR are as given in Algorithm 3 and Algorithm 4.
######
Algorithm 3 Pseudo codes of SSLSR
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**Input:**
EEG data matrix ${\overset{\sim}{\mathbf{\text{X}}}}^{\overline{N} \times C}$ and class label indicator matrix ${\overset{\sim}{\mathbf{\text{Y}}}}^{\overline{N} \times L}$, number of selected channels $\overline{C}$ and number of maximum iteration *MI*.
**Output:**
Selected feature index set 𝔽.
1: Initialize 𝔽^0^ = \[ \] **W**^0^ = **0**^*C*×*L*^, *iter* = 0.
2: **repeat**
3: *% forward*
4: Find $\left\lbrack {i^{iter},\theta^{iter}} \right\rbrack = \operatorname{argmin}\limits_{i \notin \mathbb{F}^{iter}}\operatorname{argmin}\limits_{\theta}\ L{({\mathbf{\text{W}}^{iter} + \mathbf{\text{e}}_{i}\theta^{T}})}$ by **Algorithm** **4**.
5: Update 𝔽^*iter*+1^ = 𝔽^*iter*^∪{*i*^*iter*^}.
6: Let $\mathbf{\text{W}}^{iter + 1} = \mathbf{\text{W}}^{iter} + \mathbf{\text{e}}_{i^{iter}}\theta^{iterT}$.
7: Compute $\Delta_{-} = L{(\mathbf{\text{W}}^{iter})} - L{(\mathbf{\text{W}}^{iter + 1})}$.
8: **if** Δ~−~ ≤ ϵ **then**
9: **break**
10: **end if**
11: Set *iter* = *iter*+1.
12: *% backward*
13: Find $j^{iter} = \operatorname{argmin}\limits_{j \in \mathbb{F}^{iter}}\ L{({\mathbf{\text{W}}^{iter} - \mathbf{\text{e}}_{j}\mathbf{\text{W}}_{({j,:})}^{iter}})}$.
14: Compute $\Delta_{+} = L{({\mathbf{\text{W}}^{iter} - \mathbf{\text{e}}_{j^{iter}}\mathbf{\text{W}}_{({j^{iter},:})}^{iter}})} - L{(\mathbf{\text{W}}^{iter})}$.
15: **if** Δ~+~≥νΔ~−~ **then**
16: **continue**.
17: **end if**
18: Set *iter* = *iter*−1
19: Update 𝔽^*iter*^ = 𝔽^*iter*+1^−{*j*^*iter*+1^}
20: Let $\mathbf{\text{W}}^{iter} = \mathbf{\text{W}}^{iter + 1} - \mathbf{\text{e}}_{j^{iter + 1}}\mathbf{\text{W}}_{({j^{iter + 1},:})}^{iter}$
21: **until** $\left| \mathbb{F} \middle| = \overline{N} \right.$ or *iter* = *MI*
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
######
Algorithm 4 Pseudo codes of the forward step of SSLSR
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**Input:**
EEG data matrices ${\overset{\sim}{\mathbf{\text{X}}}}^{\overline{N} \times C}$ and class label indicator matrix ${\overset{\sim}{\mathbf{\text{Y}}}}^{\overline{N} \times L}$, selected channel set 𝔽^*iter*^ and channel weight matrix **W**^*iter*^.
**Output:**
*i*^*iter*^ and θ^*iter*^.
1: Initialize $L_{min} = + \infty$ and $\mathbf{\text{T}} = \mathbf{\text{Y}} - \mathbf{\text{F}}^{iter} = \mathbf{\text{Y}} - \sum\limits_{i \in \mathbb{F}^{iter}}{\overset{\sim}{\mathbf{\text{X}}}}_{({:,i})}\mathbf{\text{W}}_{({i,:})}^{iter}$
2: **for all** *i* ∉ 𝔽^*iter*^ **do**
3: Calculate $\theta = \frac{1}{\left| \middle| {\overset{\sim}{\mathbf{\text{X}}}}_{({:,i})} \middle| |_{2}^{2} \right.}\mathbf{\text{T}}^{T}{\overset{\sim}{\mathbf{\text{X}}}}_{({:,i})}$
4: **if** ($L{({\mathbf{\text{W}}^{iter} + \mathbf{\text{e}}_{i}\theta^{T}})} < L_{min}$) **then**
5: Update $L_{min} = L{({\mathbf{\text{W}}^{iter} + \mathbf{\text{e}}_{i}\theta^{T}})}$
6: Set *i*^*iter*^ = *i* and θ^*iter*^ = θ
7: **end if**
8: **end for**
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
The theoretical proof of the convergence of Algorithm 3 can be found in Han et al. ([@B8]).
### 3.3.5. Pseudo codes of the framework
The flowchart and pseudo codes of the proposed framework are summarized in Figure [1](#F1){ref-type="fig"} and Algorithm 5.
######
Algorithm 5 Pseudo codes of the proposed FCCR
------------------------------------------------------------------------------------------------------------------------------------
**Input:**
*N* EEG data matrices $X_{1}^{M \times C}$, $X_{2}^{M \times C}$,..., and $X_{N}^{M \times C}$, hyper-parameter α, β, γ, η, ν.
**Output:**
*N*-dimensional predicted label vector $\overset{\sim}{\mathbf{\text{y}}}$.
1: EEG data pre-processing.
2: EEG feature extraction channel-wise according to Section 3.3.2.
3: Partition all trials into training and testing sets.
4: Cluster features of training trials using *k*-means.
5: EEG data representation.
6: Channel ranking and selection by **Algorithm** **1**, **Algorithm** **2**, or **Algorithm** **3**.
7: Train LDA classifier using training trials with selected channels.
8: Predict the label vector of testing trails using trained LDA.
------------------------------------------------------------------------------------------------------------------------------------
4. Results {#s4}
==========
We refer to our methods based on RFS, RUFS, and SSLSR as FCCR1, FCCR2, and FCCR3, respectively, and compare them with some baselines. In this section, we present the experimental setup, followed by the classification results.
4.1. Experimental setup
-----------------------
In this section, the baselines, metrics and other experimental settings are given in sequence.
### 4.1.1. Baselines
To validate the effectiveness of FCCRs, we compare them with the following baselines, including three kinds of static channel subset and three kinds of channel selection with training algorithms.
1. **All channels**: Signals of all available channels are used for EEG classification;
2. **3C channel s**: Signals of C3, Cz, and C4 are used for EEG classification;
3. **MEMD+STFT** (Bashar et al., [@B2]): Features extracted based on multivariate empirical mode decomposition and short time Fourier transform are used for EEG classification;
4. **gsBLDA** (Yu et al., [@B40]): Signals of channels selected based on group sparse Bayesian linear discriminant analysis are used for EEG classification;
5. **MRCS** (Zhang et al., [@B42]): Signals of channels selected by combining ReliefF and SVM are used for EEG classification;
6. **NSGA-II** (Kee et al., [@B16]): Signals of channels selected by a multi-objective genetic algorithm, i.e., NSGA-II, are used for EEG classification.
### 4.1.2. Metrics
Following previous researches, we employ the classification accuracy and F1 score to evaluate the performance of these methods.
Accuracy discovers the one-to-one relationship between predicted and ground truth labels. Denoting ${\overset{\sim}{\mathbf{\text{y}}}}_{i}$ and **y**~*i*~ as the predicted result and the ground truth label of a trial **X**~*i*~ respectively, we can compute accuracy as follows.
A
c
c
u
r
a
c
y
=
1
n
∑
i
=
1
n
δ
(
y
i
,
y
\~
i
)
×
100
\%
,
where *n* is the total number of trials and δ(*x, y*) is the delta function that equals 1 if *x* = *y* and 0 otherwise. A larger accuracy indicates a better performance.
F1 score is another widely used measure in evaluating the performance of classification. It is the harmonic mean of precision and recall,
F
1
=
2
·
P
r
e
c
i
s
i
o
n
×
R
e
c
a
l
l
P
r
e
c
i
s
i
o
n
\+
R
e
c
a
l
l
×
100
\%
.
Again, a larger F1 indicates a better performance.
### 4.1.3. Other settings
It is universally acknowledged that the performance of most algorithms is dependent on the hyper-parameters they contain. Therefore, we set some parameters in advance.
In the feature extraction process, for all methods except for MEMD+STFT, the SP features are extracted using the BCI2000 software tools (<http://www.bci2000.org>), with parameters as shown in Table [2](#T2){ref-type="table"} according to Krusienski et al. ([@B17]), McFarland and Wolpaw ([@B24]). Three TDPS, i.e., *TDP*^(0)^, *TDP*^(1)^, and *TDP*^(2)^ are extracted in this study. For MEMD+STFT, the hyper-parameters we used are identical to the original work (Bashar et al., [@B2]).
######
Parameters used to extract SP features by BCI2000.
**Model order** **First bin center** **Last bin center** **Bin width** **Evaluations per bin**
----------------- ---------------------- --------------------- --------------- -------------------------
16 0 Hz 30 Hz 3 Hz 15
In the feature clustering process, for our FCCRs, the number of clusters is determined using the \"grid-search\" strategy, i.e., from 2 to 8 with a step of 2, 10 to 90 with a step of 10, and 100 to 700 with a step of 200 on each subject, using the litekmeans algorithm.
In the channel ranking process, for gsBLDA, the stop criterion is whether the max iteration beyond 10 or the change of α or β is smaller than 0.01. For MRCS, the number of nearest neighbors *k* = 10, the maximum iteration number 50, ϵ = 0.01, and 10-fold cross-validation is used to evaluate the average classification accuracy when the top-n channels are used. For NSGA-II, the population is 300, the maximum generation is 200, and the 10-fold cross-validation accuracy is also utilized to estimate the first objective. The only difference from Kee et al. ([@B16]) and Zhang et al. ([@B42]) is that LDA, instead of the library for support vector machines (LIBSVM), is employed inside the channel selection procedure to maintain consistence with the following classification process.
In the trial classification process, 5 × 5 cross-validation on the LDA classifier is used throughout this study \[except for MEMD+STFT, in which we use k nearest neighbor (kNN) according to Bashar et al., [@B2]\]; i.e., five times for each dataset, we partition all the trials into five sets and choose four of them as the training sets, with the last set as the testing set. It should be noted that the training and testing sets are identical for all methods. For gsBLDA, MRCS, FCCR1, FCCR2, and FCCR3, the numbers of selected channels range from 1 to 118 (DS1) or 59 (DS2) with an incremental step of 1, and the numbers of the channels selected in NSGA-II are set according to the population with the best fitness.
4.2. Experimental results
-------------------------
To examine the effectiveness of the proposed framework, the classification experimental results are given and analyzed in this section. Because of the limited length of this paper, for some experiments related to subjects and features, we randomly partitioned the results for all subjects into different experiments, and show them in Figures [3](#F3){ref-type="fig"}--[5](#F5){ref-type="fig"}. Thus, all nine subjects and two features are covered, if we consider these experimental results together.
### 4.2.1. Classification performance comparison
We summarize the classification results of different methods using the two real world MI-EEG datasets in Tables [3](#T3){ref-type="table"}--[6](#T6){ref-type="table"}.
######
The mean classification accuracy (%) of nine methods on DS1.
**With the SP feature** **aa** **al** **av** **aw** **ay** **mean**
-------------------------- --------------------------- --------------------------- --------------------------- --------------------------- --------------------------- -----------
All channels 71.43 ± 4.37 83.57 ± 5.56 65.71 ± 8.41 76.07 ± 4.11 78.57 ± 3.34 75.07
3C channels 69.29 ± 5.84 **91.07** ± **2.19** 58.93 ± 3.79 79.29 ± 2.99 85.36 ± 6.61 76.79
MEMD + STFT 50.36 ± 5.11 54.64 ± 2.04 50.36 ± 5.42 39.29 ± 6.06 53.93 ± 8.32 49.71
gsBLDA 71.79(83)±4.26 87.86(42)±3.43 65.71(115)±7.08 77.86(91)±5.73 80.36(112)±4.55 76.71
MRCS 75.00(94)±1.26 84.64(89)±5.45 66.07(115)±7.14 77.50(94)±2.71 79.64(86)±7.21 76.57
NSGA-II 75.00(58)±5.50 86.43(55)±5.30 68.57(55)±6.87 80.36(58)±3.79 81.79(36)±5.56 78.43
FCCR1 75.36(82)±3.79 88.21(43)±5.14 **70.36 (11)** ± **9.57** 79.64(82)±1.60 86.43(4)±4.48 80.00
FCCR2 **80.00 (5)** ± **6.39** 90.71(80)±1.26 69.64(85)±4.07 **80.36 (97)** ± **4.82** **88.93 (4)** ± **5.14** **81.93**
FCCR3 76.79(7)±3.79 86.79(66)±2.04 68.57(72)±7.10 80.00(79)±3.87 86.79(4)±4.11 79.79
**With the TDP feature** **aa** **al** **av** **aw** **ay** **mean**
All channels 60.71 ± 3.57 85.00 ± 3.70 62.86 ± 7.19 68.21 ± 6.61 73.57 ± 3.87 70.07
3C channels 66.07 ± 10.02 91.43 ± 2.65 62.86 ± 3.87 78.21 ± 4.26 88.21 ± 5.73 77.36
MEMD + STFT 50.36 ± 5.11 54.64 ± 2.04 50.36 ± 5.42 39.29 ± 6.06 53.93 ± 8.32 49.71
gsBLDA 71.07(19)±5.56 94.29(17)±2.93 68.57(22)±5.14 80.36(14)±5.21 87.86(21)±6.11 80.43
MRCS 64.64(113)±2.93 87.86(38)±7.30 65.00(6)±5.14 76.43(20)±5.56 81.07(33)±6.00 75.00
NSGA-II 69.29(34)±6.61 94.64(26)±2.53 64.29(37)±4.19 84.64(33)±5.14 88.57(27)±4.82 80.29
FCCR1 75.00(15)±5.59 93.93(19)±3.91 **69.64 (14)** ± **5.30** 84.64(33)±6.36 88.57(26)±3.19 82.36
FCCR2 **75.71 (11)** ± **7.32** **95.36 (14)** ± **2.71** 68.21(48)±4.79 80.71(32)±4.89 90.00(20)±3.43 82.00
FCCR3 73.57(13)±7.72 94.64(14)±1.79 67.86(9)±7.14 **85.00 (29)** ± **5.59** **91.79 (10)** ± **4.11** **82.57**
*The number of selected channels (listed in the bracket) and the standard deviation are also given. The best performance is highlighted in bold*.
######
The mean classification accuracy (%) of nine methods on DS2.
**With the SP feature** **a** **b** **f** **g** **mean**
-------------------------- --------------------------- --------------------------- -------------------------- --------------------------- -----------
All channels 74.50 ± 3.26 67.00 ± 4.81 62.00 ± 4.81 72.00 ± 5.97 68.88
3C channels 69.00 ± 6.02 61.00 ± 3.79 65.00 ± 5.59 **77.50** ± **3.95** 68.13
MEMD + STFT 54.00 ± 4.18 51.00 ± 7.42 52.00 ± 5.70 49.00 ± 8.40 51.50
gsBLDA 74.50(59)±3.26 67.00(57)±6.22 70.50(2)±6.94 75.50(24)±3.26 71.88
MRCS 74.50(59)±3.26 67.00(59)±4.81 63.00(58)±4.11 76.00(36)±2.85 70.13
NSGA-II 74.50(25)±7.79 62.00(23)±2.74 68.00(9)±3.26 73.00(23)±3.26 69.38
FCCR1 76.50(7)±5.86 69.50(47)±3.06 72.00(5)±7.79 **77.50 (5)** ± **5.76** 73.87
FCCR2 **77.00 (5)** ± **5.76** **71.00 (21)** ± **8.18** **76.50 (6)** ± **9.62** 77.00(25)±6.85 **75.38**
FCCR3 76.50(42)±8.40 69.00(51)±7.20 70.00(7)±3.95 77.00(33)±6.47 73.13
**With the TDP feature** **a** **b** **f** **g** **mean**
All channels 62.50 ± 10.90 60.00 ± 7.91 55.00 ± 9.84 60.50 ± 11.51 59.50
3C channels 71.00 ± 8.02 63.50 ± 5.76 68.00 ± 6.71 80.00 ± 1.77 70.63
MEMD + STFT 54.00 ± 4.18 51.00 ± 7.42 52.00 ± 5.70 49.00 ± 8.40 51.50
gsBLDA 80.50(16)±2.74 65.00(14)±14.47 74.00(11)±8.22 77.50(17)±6.37 74.25
MRCS 76.50(25)±8.22 66.50(15)±9.29 72.50(8)±7.07 73.00(26)±6.22 72.13
NSGA-II 78.50(12)±5.76 69.00(24)±7.62 78.00(11)±5.70 79.50(10)±8.37 76.25
FCCR1 **83.50 (14)** ± **2.85** 69.50(8)±12.94 78.50(22)±3.79 **83.50 (18)** ± **4.87** 78.75
FCCR2 81.00(29)±4.68 **72.50 (9)** ± **3.26** **81.00 (9)** ± **4.81** 82.50(8)±7.20 **79.25**
FCCR3 81.00(16)±5.18 71.00(18)±6.98 80.50(21)±4.47 82.00(18)±5.42 78.63
*The number of selected channels (listed in the bracket) and the standard deviation are also given. The best performance is highlighted in bold*.
######
The mean F1-score (%) of nine methods on DS1.
**With the SP feature** **aa** **al** **av** **aw** **ay** **mean**
-------------------------- --------------------------- --------------------------- --------------------------- --------------------------- --------------------------- -----------
All channels 72.41 ± 4.15 84.25 ± 5.36 65.64 ± 8.71 75.74 ± 4.36 78.91 ± 3.90 75.39
3C channels 68.32 ± 5.69 90.71 ± 2.25 58.42 ± 3.17 77.36 ± 3.91 84.30 ± 7.87 75.82
MEMD + STFT 44.89 ± 8.77 52.84 ± 5.08 52.31 ± 3.87 36.65 ± 8.26 48.04 ± 8.95 46.94
gsBLDA 73.18(95)±3.14 88.27(39)±2.53 65.91(115)±7.16 78.45(91)±5.25 80.79(112)±5.20 77.32
MRCS 75.43(94)±2.31 85.49(87)±6.29 65.83(117)±8.22 77.30(94)±3.61 80.53(86)±6.62 76.92
NSGA-II 75.17(47)±7.88 86.97(55)±4.97 **70.06 (55)** ± **6.54** **80.73 (58)** ± **4.12** 81.50(36)±5.79 78.89
FCCR1 76.16(113)±4.69 88.61(43)±4.56 69.93(11)±8.90 79.86(82)±2.97 87.16(4)±7.20 80.34
FCCR2 **80.99 (5)** ± **5.68** **91.18 (80)** ± **1.03** 69.66(60)±5.35 80.51(97)±4.85 **89.53 (4)** ± **3.59** **82.37**
FCCR3 78.44(7)±3.06 87.55(66)±1.82 68.61(55)±4.04 79.65(79)±4.34 87.11(4)±4.00 80.27
**With the TDP feature** **aa** **al** **av** **aw** **ay** **mean**
All channels 58.52 ± 3.89 85.09 ± 3.75 62.61 ± 6.21 68.59 ± 7.62 73.29 ± 5.53 69.62
3C channels 66.94 ± 8.42 91.70 ± 2.42 63.49 ± 4.82 77.82 ± 6.03 88.76 ± 4.84 77.74
MEMD + STFT 44.89 ± 8.77 52.84 ± 5.08 52.31 ± 3.87 36.65 ± 8.26 48.04 ± 8.95 46.94
gsBLDA 71.89(13)±4.91 94.49(17)±2.66 68.51(22)±5.54 79.75(54)±6.53 87.92(21)±6.28 80.51
MRCS 64.09(113)±4.34 88.46(38)±6.40 67.62(6)±6.83 76.05(20)±5.81 81.49(33)±5.64 75.54
NSGA-II 69.36(34)±6.10 94.68(26)±2.48 65.21(28)±10.91 **84.91 (33)** ± **4.65** 88.73(27)±4.59 80.58
FCCR1 75.86(15)±4.86 94.09(33)±3.74 **69.86 (14)** ± **3.96** 84.67(33)±4.88 88.68(26)±1.48 **82.63**
FCCR2 **76.42 (11)** ± **6.31** **95.55 (18)** ± **4.98** 69.60(48)±7.48 80.71(35)±4.62 90.00(20)±5.27 82.45
FCCR3 73.30(13)±8.48 94.85(32)±3.42 68.45(7)±3.79 84.57(29)±5.79 **91.77 (10)** ± **3.99** 82.59
*The number of selected channels (listed in the bracket) and the standard deviation are also given. The best performance is highlighted in bold*.
######
The mean F1-score (%) of nine methods on DS2.
**With the SP feature** **a** **b** **f** **g** **mean**
-------------------------- --------------------------- --------------------------- --------------------------- --------------------------- -----------
All channels 74.09 ± 2.60 67.95 ± 4.34 60.77 ± 8.73 71.51 ± 7.26 68.58
3C channels 66.70 ± 5.44 62.08 ± 3.37 62.65 ± 8.19 **76.57** ± **4.63** 67.00
MEMD + STFT 51.11 ± 2.99 40.23 ± 7.26 50.25 ± 9.65 47.93 ± 7.15 47.38
gsBLDA 74.09(59)±2.60 68.27(57)±5.93 69.10(2)±8.51 75.05(24)±5.18 71.63
MRCS 74.09(59)±2.60 68.24(58)±3.75 63.38(11)±5.77 75.56(36)±4.12 70.32
NSGA-II 74.15(25)±7.81 63.61(23)±3.66 66.14(9)±4.13 71.33(23)±2.53 68.81
FCCR1 75.49(7)±4.80 70.88(47)±5.29 72.59(5)±9.90 75.79(57)±7.91 73.69
FCCR2 75.85(5)±6.38 **72.44 (21)** ± **9.74** **76.12 (6)** ± **10.31** 76.11(25)±5.34 **75.13**
FCCR3 **76.17 (46)** ± **6.13** 70.35(51)±5.71 70.59(7)±3.22 76.05(24)±6.48 73.29
**With the TDP feature** **a** **b** **f** **g** **mean**
All channels 63.18 ± 9.86 58.52 ± 8.57 50.94 ± 16.99 61.42 ± 11.54 58.51
3C channels 68.87 ± 8.28 63.29 ± 5.21 67.31 ± 8.44 79.38 ± 1.24 69.71
MEMD + STFT 51.11 ± 2.99 40.23 ± 7.26 50.25 ± 9.65 47.93 ± 7.15 47.38
gsBLDA 78.99(16)±1.43 65.29(4)±5.16 74.77(11)±7.40 78.11(18)±3.32 74.29
MRCS 75.58(25)±10.68 65.94(15)±10.66 71.94(8)±8.50 70.81(23)±5.18 71.07
NSGA-II 78.29(12)±5.98 69.75(24)±7.83 78.47(11)±6.10 78.52(13)±7.37 76.26
FCCR1 **83.59 (14)** ± **3.41** 69.30(8)±14.41 78.17(22)±4.01 **82.88 (18)** ± **5.21** 78.48
FCCR2 80.09(34)±3.90 **72.81 (9)** ± **3.12** **80.90 (9)** ± **4.37** 82.14(8)±5.37 **78.99**
FCCR3 79.87(16)±5.93 71.46(18)±6.76 80.35(21)±5.57 81.76(16)±4.49 78.36
*The number of selected channels (listed in the bracket) and the standard deviation are also given. The best performance is highlighted in bold*.
From these four tables, we make the following observations. First, the classification performance of the selected channel subsets is superior to All channels, demonstrating the practical significance of channel selection. For instance, as shown in Table [3](#T3){ref-type="table"}, the best mean accuracies for the selected channels on DS1 are 81.93 and 82.57%; i.e., they are 9.14 and 17.8% better than the 75.07 and 70.07% for All channels, respectively.
Second, channel subsets selected through training often perform better than static channel subsets, as they verify the role of channel selection approaches using training. Taking Table [6](#T6){ref-type="table"} as an example, the best mean F1 scores on DS2 for training methods are 75.13 and 78.99%, which outperform the 67.00 and 69.71% of the static 3C channels by 12.1 and 13.3%, respectively.
Third, our proposed FCCRs usually outperform other channel selection methods on various subjects from different datasets. In terms of mean accuracies and F1 scores, our FCCRs consistently obtain the best performance. Specifically, of 36 results on nine subjects with two features in terms of two metrics, the FCCRs gain the 31 best results, occupying 86.1%.
Finally, our FCCRs can sometimes gain the best performance with fewer channels, indicating the effectiveness of our framework. Especially on subject ay with SP features, all three FCCRs obtain high accuracies and F1 scores with only four channels. In contrast, the fewest channels selected by other methods are 36 (NSGA-II), nine times the FCCRs\' results.
The *p*-values of one-way analysis of variance (ANOVA)[^1^](#fn0001){ref-type="fn"} statistical tests of the improvement for all methods are plotted in Figure [2](#F2){ref-type="fig"}, from which we can observe that the FCCRs obtain a significant improvement over the other methods in most situations. The three FCCRs have 18 improvements over six baselines for each panel; an average of 14 improvements is significant, occupying 77.8%.
{#F2}
Except for the above, we plot the precision and recall (P-R) pairs for all methods in Figure [3](#F3){ref-type="fig"}, from which we can obtain other observations. First, some channel selection approaches may be limited in the recall, e.g., MRCS and NSGA-II on subject aa with SP features; gsBLDA, MRCS and NSGA-II on subject aw with TDP features; and gsBLDA on subject a with TDP features.
{#F3}
Second, our FCCRs can obtain a relatively high precision with acceptable recall. Specifically, on subject aw with TDP features, FCCR2, and FCCR3 both gain the recall of 1 and the precision larger than 0.8 at the same time, much better than the baselines.
Third, the P-R curve varies obviously with the subjects. For subjects aa and b with SP features, the recall of all the methods is below 0.9, and the precision is hardly beyond 0.9. Conversely, many markers are distributed in the area beyond 0.9 in terms of both precision and recall for subjects aw and a with TDP features.
### 4.2.2. Selected channels comparison
To visually compare the performance of the different channel selection approaches, we plot the classification accuracies of nine methods with different numbers of selected channels on four subjects in Figure [4](#F4){ref-type="fig"}. The observations are as follows. First, almost all channel selection methods are effective, including 3C channels, gsBLDA, MRCS, NSGA-II and three FCCRs, since their accuracies with fewer selected channels can outperform the ones with all channels. Second, few channel selection results are superior to the ones with 3C channels. In fact, on subject g with TDP features, only the FCCRs outperform the 3C channels. Third, the proposed FCCRs can obtain the best performance compared with the other methods in most situations. For these four results, the best performance is always acquired by FCCRs. Fourth, our FCCRs are able to gain the highest accuracies with few channels, which is consistent with the above results.
{#F4}
Additionally, in Figure [5](#F5){ref-type="fig"}, we display the selected channel subset on the subject\'s scalp with the top classification accuracies for subject ay, in which the channels ranked from top to bottom are indicated by the red to blue areas. It is clear that significant and informative channels relevant to the MI-EEG, i.e., channels in the motor-sensory cortex, are selected first by the proposed FCCRs. gsBLDA, MRCS and NSGA-II usually select some irrelevant or redundant channels, possibly having an adverse effect on their classification performance, which is consistent with the results in section 4.2.1 and Figure [4](#F4){ref-type="fig"}.
{#F5}
### 4.2.3. Computational complexity comparison
In this subsection, we experimentally analyze the computational complexity of nine methods by comparing their average running time (ART) for nine subjects with two datasets, on a PC with a 4.0GHz Intel i-7 6700K CPU with 32 GB of memory.
From Figure [6](#F6){ref-type="fig"} and Table [7](#T7){ref-type="table"}, we can observe that the ARTs of the FCCRs are longer than some approaches without training, including All channels and 3C channels. In addition, the fact that the FCCRs run faster than MRCS, NSGA-II, and MEMD+STFT demonstrates their efficiency. It is interesting that MEMD+STFT is unexpectedly slow, even though it has no channel selection. It is also clear that the ARTs of FCCRs are \< 10 s on nine subjects, which is comparable with motor imagery paradigms, demonstrating their practical worth in real applications.
{#F6}
######
Average running time (ART) measured in *s* of nine methods on two datasets.
**Method** **DS1** **DS2** **Mean**
-------------- ----------- ----------- ----------- ----------- -----------
All channels 1.086 0.253 0.685 0.153 0.544
3C channels **1.031** **0.242** **0.637** **0.149** **0.515**
MEMD + STFT 13,949 13,949 10,752 10,752 12,351
gsBLDA 1.670 0.354 0.818 0.196 0.760
MRCS 43.31 17.78 71.42 37.76 42.57
NSGA-II 8,832 3,316 5,368 2,790 5,077
Proposed1 6.828 2.149 1.764 1.105 2.962
Proposed2 6.645 6.394 1.811 0.858 3.927
Proposed3 8.100 1.961 1.187 1.580 3.207
*The shortest ARTs are highlighted in bold*.
### 4.2.4. Parameter sensitivity
The insensitivity of our proposed framework to cluster numbers is demonstrated in Figure [7](#F7){ref-type="fig"}, in which we notice that cluster numbers exert a limited influence on classification accuracies in quite a large range, from 2 to 300.
{#F7}
### 4.2.5. Comprehensive comparison
In addition to the above experiments, we comprehensively compare nine methods comprehensively in terms of six metrics, i.e., mean accuracy (MA), mean F1 score (MF), mean precision (MP), mean recall (MR), average running speed (ARS, number of trials divided by ART), and average channel reduction ratio (ACRR, ratio of number of removed channels to all channels), and show the results in Figure [8](#F8){ref-type="fig"}. In this figure, the inner hexagon indicates 0 for MA, MF, MP, and MR, the smallest ARS, and 0% for ACRR. The outer hexagon indicates 1 for MA, MF, MP, and MR, the largest ARS, and 100% for ACRR. Thus, for all results, more outside markers mean better performance.
![Mean accuracy, mean F1 score, mean precision, mean recall, average running speed and average channel reduction ratio for nine algorithms. Except for the average running speed, whose range varies according to the practical minimum and maximum values of the nine methods, the range of the other metrics is \[0,1\] (\[0%,100%\] for ACRR).](fnins-12-00217-g0008){#F8}
It appears from Figure [8](#F8){ref-type="fig"} that, for all data sets with SP or TDP features, the FCCRs are superior to other state-of-the-art methods in terms of almost all metrics, except for ARS.
5. Discussion {#s5}
=============
From the experimental results shown above, we find that our methods are conducive to EEG classification through feature compression, low-dimensional data representation, and convergent channel ranking. The primary reasons for the promising classification performance of our FCCRs include the following: (1) motor-sensory cortex signals are much more relevant to MI-EEG than signals of other cortexes; signals from redundant channels corresponding to unrelated cortexes may negatively influence the EEG classification performance; (2) informative channels often vary between subjects and the optimal channel subsets need to be found empirically; the intrinsic physiological and mental condition divergence among subjects during the experimental process is fairly large; (3) our FCCRs can probably select more pivotal channels and rank them in the top position, by virtue of their channel selection methods with high capability and strict convergence.
With regard to the selected channel comparison results, the primary explanations follow: (1) signals from irrelevant and redundant channels have a negative effect on classifying MI-EEG trials, as they introduce confusing information; (2) 3C channels, i.e., Cz, C3, and C4, are closely related to the motor-sensory cortex and pivotal for classifying motor imagery EEG signals; they are infrequently selected by some methods.
The computational complexity comparison results could be intuitively foreseen by comparing their flowcharts and procedures. The detailed explanation mainly includes the following: (1) the process of evaluating the channels\' weights is unnecessary in methods without training, e.g., All channels and 3C channels; thus, their ARTs are small; (2) the FCCRs compress the feature vectors and exploit strictly convergent channel selection approaches to reduce the data scale and the period of evaluating the channels\' weights; they thereby accelerate the EEG classification procedure and decrease the FCCRs\' ARTs; (3) methods with embedded classifiers, e.g., MRCS and NSGA-II, need to estimate the channel subset and update the channel\'s weights in each iteration, leading to large ARTs. Additionally, MEMD+STFT has large ARTs because extracting the MEMD features is time consuming, especially when the total number of signal projections is beyond 200 in the MEMD algorithm.
For the parameter sensitivity results, the cluster numbers have limited influence on estimating the similarity among features and clustering-based representative signals. Therefore, changing the number of clusters does not significantly affect the following channel selection and classification.
From the comprehensive comparison results, we find that the FCCRs have some exciting advantages, including a small data scale, a strictly convergent procedure, short running time, and high classification accuracy. All of these visibly promote their comprehensive performance, which is helpful for implementing them with field programmable gate arrays (FPGAs) or digital signal processors (DSPs) in practical applications.
As shown in Figure [1](#F1){ref-type="fig"}, the main differences between the FCCRs and previous studies are feature compression and convergent channel ranking. On one hand, in current EEG feature extraction methods, spatial features, or features in the time and frequency domain, are always *directly* used to select channels, and even classify trials (Blankertz et al., [@B5]; Zhaoyang et al., [@B43]). From the perspective of classical pattern recognition, the essential role of the features depends on their similarities or differences, rather than the explicit numerical values themselves. Therefore, compressing features by clustering them and endowing them with signatures not only preserves the intrinsic similarities involved in the features, but also reduces the data scale. On the other hand, most existing channel selection methods carefully evaluate the channels\' weights by *directly* classifying the signals that compose them (Kee et al., [@B16]; Zhang et al., [@B42]). Thus, their performance is classifier-dependent, lacking adaptability and support for unfamiliar situations. In most practical applications, classifiers are not fixed and should be flexibly chosen according to practical conditions. Moreover, ignoring the convergence of the selection methods may lead to unsatisfactory results, since the process of adding or removing channels is not controllable. Thus, it is better to leverage classifier-independent and performance-verified selection approaches with proven strict convergence to guarantee both the process and the result of the channel weights evaluation. Philosophically, some solutions have tortuous exteriors but are very rapid in many researching areas. By introducing feature compression and convergent channel selection, we propose the FCCRs as such a promising opportunity to classify multiple EEG signal types.
Inevitably, FCCRs still have some potential limitations to be overcome, which may bring a forth series of innovations and applications by integrating plenty of current and forthcoming signal processing and pattern recognition approaches in the future. Firstly, we will focus on the feature extraction and clustering methods for seizing the key information from the original EEG signals with low SNRs. Moreover, considering the ever-developing feature selection approaches, dynamic weighting techniques for evaluating channels are also among our directions for future research. In addition, deciding the optimal number of channels is still an open problem, for which we will attempt to propose a criterion in the near future.
6. Conclusion {#s6}
=============
In this paper, we present a novel EEG classification framework named FCCR, which introduces feature compression and normal signal representation between the traditional feature extraction and channel selection procedures. In addition, some strict convergent algorithms with theoretical proofs are leveraged to rank and select informative channels, thereby showing that our proposed framework is capable of classifying EEG signals with acceptable accuracy and speed. Extensive experimental results on two real world motor imagery EEG datasets validate the effectiveness and efficiency of the proposed framework, as compared with state-of-the-art methods.
Author contributions {#s7}
====================
JH, JZ, and CW designed research. JH and YZ performed research. AK, GX, and HZ analyzed data and drafted parts of the first version of the manuscript. JH, HS, and JC wrote and revised the paper under the supervision of JZ and CW.
Conflict of interest statement
------------------------------
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
^1^To completely compare the methods with channel selection, we compute the *p*-values using the Acc or F1 across all 118(DS1) or 59(DS2) channels and five validations for gsBLDA, MRCS and the FCCRs. The results in Tables [3](#T3){ref-type="table"}--[6](#T6){ref-type="table"} are used to compute the *p*-values for other methods.
**Funding.** This work is financially supported by International Cooperation and Exchange of the National Natural Science Foundation of China (No. 31320103914), General Program of National Natural Science Foundation of China (No. 31370987), National Natural Science Funds for Outstanding Young Scholar (No. 81622027), Beijing NOVA Program of China (No.2016B615), and National Key Research and Development Program of China (No.2017YFA0106100).
For any vector **x** ∈ ℝ^*n*×1^, the *l*~*p*~-norm is defined as
\|
\|
x
\|
\|
p
=
(
∑
i
=
1
n
\|
x
i
\|
p
)
1
p
,
where *x*~*i*~ is the *i*-th element of **x**. For any matrix **X** ∈ ℝ^*n*×*m*^, the *m*~*r*~ -norm and *m*~*r,\ s*~ -norm are defined as
\|
\|
X
\|
\|
r
=
(
∑
i
=
1
n
∑
j
=
1
m
\|
X
i
j
\|
r
)
1
r
,
\|
\|
X
\|
\|
r
,
s
=
(
∑
i
=
1
n
(
∑
j
=
1
m
\|
X
i
j
\|
r
)
s
r
)
1
s
.
[^1]: Edited by: Mikhail Lebedev, Duke University, United States
[^2]: Reviewed by: Zhong Yin, University of Shanghai for Science and Technology, China; Yufeng Ke, Tianjin University, China
[^3]: This article was submitted to Neural Technology, a section of the journal Frontiers in Neuroscience
|
// Copyright (c) 2017 Computer Vision Center (CVC) at the Universitat Autonoma
// de Barcelona (UAB).
//
// This work is licensed under the terms of the MIT license.
// For a copy, see <https://opensource.org/licenses/MIT>.
#include <carla/PythonUtil.h>
#include <carla/rpc/Command.h>
#include <carla/rpc/CommandResponse.h>
#define TM_DEFAULT_PORT 8000
namespace command_impl {
template <typename T>
const T &Convert(const T &in) {
return in;
}
carla::rpc::ActorId Convert(const boost::shared_ptr<carla::client::Actor> &actor) {
return actor->GetId();
}
carla::rpc::ActorDescription Convert(const carla::client::ActorBlueprint &blueprint) {
return blueprint.MakeActorDescription();
}
template <typename... ArgsT>
static boost::python::object CustomInit(boost::python::object self, ArgsT... args) {
return self.attr("__init__")(Convert(args)...);
}
template <typename... ArgsT>
static boost::python::object CustomSpawnActorInit(boost::python::object self, ArgsT... args) {
return self.attr("__init__")(carla::rpc::Command::SpawnActor{Convert(args)...});
}
static carla::rpc::Command::SpawnActor Then(
carla::rpc::Command::SpawnActor &self,
carla::rpc::Command command) {
self.do_after.push_back(command);
return self;
}
} // namespace command_impl
void export_commands() {
using namespace boost::python;
namespace cc = carla::client;
namespace cg = carla::geom;
namespace cr = carla::rpc;
using ActorPtr = carla::SharedPtr<cc::Actor>;
object command_module(handle<>(borrowed(PyImport_AddModule("libcarla.command"))));
scope().attr("command") = command_module;
scope submodule_scope = command_module;
// This is a handler for passing to "SpawnActor.then" commands.
submodule_scope.attr("FutureActor") = 0u;
class_<cr::CommandResponse>("Response", no_init)
.add_property("actor_id", +[](const cr::CommandResponse &self) {
return self.HasError() ? 0u : self.Get();
})
.add_property("error", +[](const cr::CommandResponse &self) {
return self.HasError() ? self.GetError().What() : std::string("");
})
.def("has_error", &cr::CommandResponse::HasError)
;
class_<cr::Command::SpawnActor>("SpawnActor")
.def(
"__init__",
&command_impl::CustomSpawnActorInit<cc::ActorBlueprint, cg::Transform>,
(arg("blueprint"), arg("transform")))
.def(
"__init__",
&command_impl::CustomSpawnActorInit<cc::ActorBlueprint, cg::Transform, const cr::ActorId &>,
(arg("blueprint"), arg("transform"), arg("parent_id")))
.def(
"__init__",
&command_impl::CustomSpawnActorInit<cc::ActorBlueprint, cg::Transform, ActorPtr>,
(arg("blueprint"), arg("transform"), arg("parent")))
.def(init<cr::Command::SpawnActor>())
.def_readwrite("transform", &cr::Command::SpawnActor::transform)
.def_readwrite("parent_id", &cr::Command::SpawnActor::parent)
.def("then", &command_impl::Then, (arg("command")))
;
class_<cr::Command::DestroyActor>("DestroyActor")
.def("__init__", &command_impl::CustomInit<ActorPtr>, (arg("actor")))
.def(init<cr::ActorId>((arg("actor_id"))))
.def_readwrite("actor_id", &cr::Command::DestroyActor::actor)
;
class_<cr::Command::ApplyVehicleControl>("ApplyVehicleControl")
.def("__init__", &command_impl::CustomInit<ActorPtr, cr::VehicleControl>, (arg("actor"), arg("control")))
.def(init<cr::ActorId, cr::VehicleControl>((arg("actor_id"), arg("control"))))
.def_readwrite("actor_id", &cr::Command::ApplyVehicleControl::actor)
.def_readwrite("control", &cr::Command::ApplyVehicleControl::control)
;
class_<cr::Command::ApplyWalkerControl>("ApplyWalkerControl")
.def("__init__", &command_impl::CustomInit<ActorPtr, cr::WalkerControl>, (arg("actor"), arg("control")))
.def(init<cr::ActorId, cr::WalkerControl>((arg("actor_id"), arg("control"))))
.def_readwrite("actor_id", &cr::Command::ApplyWalkerControl::actor)
.def_readwrite("control", &cr::Command::ApplyWalkerControl::control)
;
class_<cr::Command::ApplyTransform>("ApplyTransform")
.def("__init__", &command_impl::CustomInit<ActorPtr, cg::Transform>, (arg("actor"), arg("transform")))
.def(init<cr::ActorId, cg::Transform>((arg("actor_id"), arg("transform"))))
.def_readwrite("actor_id", &cr::Command::ApplyTransform::actor)
.def_readwrite("transform", &cr::Command::ApplyTransform::transform)
;
class_<cr::Command::ApplyWalkerState>("ApplyWalkerState")
.def("__init__", &command_impl::CustomInit<ActorPtr, cg::Transform, float>, (arg("actor"), arg("transform"), arg("speed")))
.def(init<cr::ActorId, cg::Transform, float>((arg("actor_id"), arg("transform"), arg("speed"))))
.def_readwrite("actor_id", &cr::Command::ApplyWalkerState::actor)
.def_readwrite("transform", &cr::Command::ApplyWalkerState::transform)
.def_readwrite("speed", &cr::Command::ApplyWalkerState::speed)
;
class_<cr::Command::ApplyTargetVelocity>("ApplyTargetVelocity")
.def("__init__", &command_impl::CustomInit<ActorPtr, cg::Vector3D>, (arg("actor"), arg("velocity")))
.def(init<cr::ActorId, cg::Vector3D>((arg("actor_id"), arg("velocity"))))
.def_readwrite("actor_id", &cr::Command::ApplyTargetVelocity::actor)
.def_readwrite("velocity", &cr::Command::ApplyTargetVelocity::velocity)
;
class_<cr::Command::ApplyTargetAngularVelocity>("ApplyTargetAngularVelocity")
.def("__init__", &command_impl::CustomInit<ActorPtr, cg::Vector3D>, (arg("actor"), arg("angular_velocity")))
.def(init<cr::ActorId, cg::Vector3D>((arg("actor_id"), arg("angular_velocity"))))
.def_readwrite("actor_id", &cr::Command::ApplyTargetAngularVelocity::actor)
.def_readwrite("angular_velocity", &cr::Command::ApplyTargetAngularVelocity::angular_velocity)
;
class_<cr::Command::ApplyImpulse>("ApplyImpulse")
.def("__init__", &command_impl::CustomInit<ActorPtr, cg::Vector3D>, (arg("actor"), arg("impulse")))
.def(init<cr::ActorId, cg::Vector3D>((arg("actor_id"), arg("impulse"))))
.def_readwrite("actor_id", &cr::Command::ApplyImpulse::actor)
.def_readwrite("impulse", &cr::Command::ApplyImpulse::impulse)
;
class_<cr::Command::ApplyForce>("ApplyForce")
.def("__init__", &command_impl::CustomInit<ActorPtr, cg::Vector3D>, (arg("actor"), arg("force")))
.def(init<cr::ActorId, cg::Vector3D>((arg("actor_id"), arg("force"))))
.def_readwrite("actor_id", &cr::Command::ApplyForce::actor)
.def_readwrite("force", &cr::Command::ApplyForce::force)
;
class_<cr::Command::ApplyAngularImpulse>("ApplyAngularImpulse")
.def("__init__", &command_impl::CustomInit<ActorPtr, cg::Vector3D>, (arg("actor"), arg("impulse")))
.def(init<cr::ActorId, cg::Vector3D>((arg("actor_id"), arg("impulse"))))
.def_readwrite("actor_id", &cr::Command::ApplyAngularImpulse::actor)
.def_readwrite("impulse", &cr::Command::ApplyAngularImpulse::impulse)
;
class_<cr::Command::ApplyTorque>("ApplyTorque")
.def("__init__", &command_impl::CustomInit<ActorPtr, cg::Vector3D>, (arg("actor"), arg("torque")))
.def(init<cr::ActorId, cg::Vector3D>((arg("actor_id"), arg("torque"))))
.def_readwrite("actor_id", &cr::Command::ApplyTorque::actor)
.def_readwrite("torque", &cr::Command::ApplyTorque::torque)
;
class_<cr::Command::SetSimulatePhysics>("SetSimulatePhysics")
.def("__init__", &command_impl::CustomInit<ActorPtr, bool>, (arg("actor"), arg("enabled")))
.def(init<cr::ActorId, bool>((arg("actor_id"), arg("enabled"))))
.def_readwrite("actor_id", &cr::Command::SetSimulatePhysics::actor)
.def_readwrite("enabled", &cr::Command::SetSimulatePhysics::enabled)
;
class_<cr::Command::SetAutopilot>("SetAutopilot")
.def("__init__", &command_impl::CustomInit<ActorPtr, bool, uint16_t>, (arg("actor"), arg("enabled"), arg("tm_port") = TM_DEFAULT_PORT ))
.def(init<cr::ActorId, bool, uint16_t>((arg("actor_id"), arg("enabled"), arg("tm_port") = TM_DEFAULT_PORT )))
.def_readwrite("actor_id", &cr::Command::SetAutopilot::actor)
.def_readwrite("tm_port", &cr::Command::SetAutopilot::tm_port)
.def_readwrite("enabled", &cr::Command::SetAutopilot::enabled)
;
class_<cr::Command::SetVehicleLightState>("SetVehicleLightState")
.def("__init__", &command_impl::CustomInit<ActorPtr, bool>, (arg("actor"), arg("light_state")))
.def(init<cr::ActorId, cr::VehicleLightState::flag_type>((arg("actor_id"), arg("light_state"))))
.def_readwrite("actor_id", &cr::Command::SetVehicleLightState::actor)
.def_readwrite("light_state", &cr::Command::SetVehicleLightState::light_state)
;
implicitly_convertible<cr::Command::SpawnActor, cr::Command>();
implicitly_convertible<cr::Command::DestroyActor, cr::Command>();
implicitly_convertible<cr::Command::ApplyVehicleControl, cr::Command>();
implicitly_convertible<cr::Command::ApplyWalkerControl, cr::Command>();
implicitly_convertible<cr::Command::ApplyTransform, cr::Command>();
implicitly_convertible<cr::Command::ApplyWalkerState, cr::Command>();
implicitly_convertible<cr::Command::ApplyTargetVelocity, cr::Command>();
implicitly_convertible<cr::Command::ApplyTargetAngularVelocity, cr::Command>();
implicitly_convertible<cr::Command::ApplyImpulse, cr::Command>();
implicitly_convertible<cr::Command::ApplyForce, cr::Command>();
implicitly_convertible<cr::Command::ApplyAngularImpulse, cr::Command>();
implicitly_convertible<cr::Command::ApplyTorque, cr::Command>();
implicitly_convertible<cr::Command::SetSimulatePhysics, cr::Command>();
implicitly_convertible<cr::Command::SetAutopilot, cr::Command>();
implicitly_convertible<cr::Command::SetVehicleLightState, cr::Command>();
}
|
By: Asif Balouch
Cross-posted from: http://philasify101.blogspot.com/2012/05/top-10-traits-of-real-man-muslim-style.html#more
A phrase that has long been thrown around is the popular term, “Real Man”. Being labeled a “Real Man” has forever been seen as the ultimate compliment of respect a man could give another and to be viewed in this way has been regarded as the ultimate goal to achieve for the male population. But what is a “Real Man”?
Support MuslimMatters for Just $2 a Month MuslimMatters has been a free service to the community since 2007. All it takes is a small gift from a reader like you to keep us going, for just $2 / month. The Prophet (SAW) has taught us the best of deeds are those that done consistently, even if they are small. Click here to support MuslimMatters with a monthly donation of $2 per month. Set it and collect blessings from Allah (swt) for the khayr you're supporting without thinking about it.
In some parts of society this achievement is measured in terms of how much alcohol you can consume without keeling over and puking your guts out. Others regard a real man in accordance to how many fights one has participated in and won, even if the reason for the fight is for something completely idiotic and dishonorable. Another historic definition of a real man would depend on how shredded and cut up he is when it comes to his muscularity, never mind if the method of gaining it was questionable. Some societies label one a “real man” depending on how much money he rakes in or how many sexual encounters he’s had, disregarding whether the money is earned in a positive manner and overlooking the tremendous heartbreak of the poor women that agreed to be another tally mark to his bedroom antics. Catching a football, driving a pickup truck…I’m sure there’s plenty more.
The other day I came across an article on AskMen.com that provided a top 10 list of traits of a real man. Now granted, I don’t think the blog was supposed to be taken too seriously because it was from the perspective of a mafioso character. Despite that, I thought it was a good concept to dive into when properly defining a “Real Man”. I believe this top 10 list is significant since what constitutes a real man is not legislated by any regular humanfolk. Oh, no, no, no…the declaration of what a Real Man comes from the words of the Creator, Allah (God) of man himself, and was put into practice by a man who is without a doubt the epitome of a real man, The Prophet Muhammad . Just by reading about this man’s life, his struggles and the description of his profound character, there is no one on this earth that the label “Real Man” fits more perfectly than him. So without further ado, here are the top ten traits of a real man.
Trait # 1: A Real Man Reads
That’s right, just like how your stomach needs to be regularly fed with food, your mind needs to be regularly fed with knowledge. Now there’s quite a few ways to attain knowledge of different things but a surefire way for it to stick in your brain is through reading. (A quick note: If you’re a real man, you didn’t groan when reading this first trait.) And no I’m not saying reading the newspaper, or your Twitter or Facebook news feeds does it. That is NOT true reading. That’s not manly reading. I’m talking about books. A man reads books. Books that can save him from making stupid mistakes, books that can inspire him to get off his lazy rear end and do something with himself, books that makes him realize why the heck he’s here in the first place. Though the Prophet himself was illiterate, he made it a high priority to seek knowledge and there are many sayings where he mentions the high value importance of attaining knowledge. Some include: “The pen(knowledge) is mightier than the sword” and “He who travels in the search of knowledge, to him God shows the way of Paradise.” A real man doesn’t stop seeking knowledge until his heart stops beating.
Trait # 2: A Real Man is a Focused Man
Alluding to the trait above, a real man who has knowledge, recognizes what is important and what is rubbish that ain’t worth his time. A real man doesn’t lose himself in the pretty bells and whistles of life that don’t bring him any real benefit. When a real man sees things that are a distraction, he keeps moving. A real man realizes why he’s here on this earth, that he is only here for a short time and that he has to do what he has to do to make sure he’s got a ticket to paradise. A real man focuses on what’s important: bettering himself, making a living for his family and preparing himself for the future. A real man lives by what the Prophet said:“Live in this world as (if you are) a wayfarer or a stranger.”, which in essence means to not get caught up by all the glitz and glamour of this world because it’s temporary. He doesn’t stick around and waste time. Don’t let things distract you because you have to make it to your destination. A real man manages his time wisely and gives the proper due to his work, his family, and his Creator. Stay focused and do what you have to do, because in the end there isn’t anybody to bail you out.
Trait # 3: A Real Man is Gentle but Firm
A real man recognizes that he’s a man with an intellect, not an animal. A real man speaks softly and doesn’t need to holler and yell at the top of his lungs to be heard or get his point across. A real man doesn’t create a scene, start trouble or “act hard” in front of others to show he’s a tough guy. When there is trouble ahead, he does his best to squash it in a civilized manner. A real man isn’t a punk either. Just because he doesn’t raise his voice or try to intimidate others, doesn’t mean he’s a floor mat that people can walk all over. If he has problem with something, he lets it be heard. A real man practices patience. A real man suppresses his anger like a caged lion. Sure he can open the door at anytime and unleash hell but knows there is more honor and dignity in holding back. Raising his fists is the ABSOLUTE last resort.The Prophet said, “He is not strong and powerful who throws people down, but he is strong who withholds himself from anger.”, in addition he also said, “Deal gently with people, and be not harsh; cheer them and condemn not.” A real man lives by these quotes.
Trait # 4: A Real Man is a Family Man
A real man recognizes the importance of his family and does his utmost to be a contributing member of the family. A real man preserves and protects. A real man recognizes that his children are a blessing from God and treats them as such and brings them up to be upright human beings. A real man MAKES TIME for his family and does not neglect them because of work or his own personal dealings. A real man is the backbone of his family and doesn’t have time to be weak. A real man doesn’t just take care of his immediate household but looks after his family that he grew up with. He honors his parents to the utmost, especially his mother. He calls his family often and stays in touch. He’s good to his siblings and relatives. A real man strives to be the best father, brother, son and husband he can be and works hard to live up to be the best. The Prophet once said, “The best of you is the one who is best to his own family, and I am the best of you towards my family.”
Trait # 5: A Real Man doesn’t Slander/Backbite/Cuss/Gossip
A real man keeps his mouth shut if he doesn’t have anything nice to say. A real man, when he hears others ripping on whoever–whether they know them or not, either goes over and shuts it down by warning the party or walks right on out of the room. A real man would say in that situation, “Hey, I wouldn’t like anyone saying that about me when I’m not around so ya’ll shouldn’t be talking about so-and-so like that”. A real man keeps a careful watch on his tongue because he knows that what he says can hurt him then or definitely in the future. A real man doesn’t discuss things he doesn’t know about or people he hasn’t ever met. A real man doesn’t cuss to be “cool” and chooses his words intelligently because he can get his point across better without dropping an f-bomb or an s-missle. A real man recognizes the hadith that says “most people that are in hellfire are in there because they couldn’t control their tongue.” In the Qurʾān, God states that he has given man “two lips and one tongue”, so the lips can control the tongue. [Surah Al Balad, 9]
Trait # 6: A Real Man Keeps His Promises
A real man’s word is his bond. If he can’t keep a promise, he doesn’t give his word. A real man is trustworthy and doesn’t flake out on somebody. He doesn’t use Insha’Allah (God willing) as a copout. He doesn’t break deals and he pays back debts. A real man knows that his words are as powerful as his actions, and that they must be taken at face value. A real man doesn’t say “I’ll try” if he doesn’t have to try it really. He either does or he doesn’t and if he can’t do it, he says he can’t. There’s no shame in saying that you can’t do a favor for somebody, or you won’t be able to come through. At least you’ll be upfront and honest about it rather than being relied on and letting someone down. What does the Qurʾān say about this? “O you who believe! Why do you say that which you do not do? Most hateful it is with Allah that you say that which you do not do.” [Surah Saff, 2-3] From this quote, a Real Man knows that keeping a promise is serious business and doesn’t screw around on them.
Trait # 7: A Real Man Respects All Women
A real man doesn’t “holler” at girls. A real man doesn’t sit with other guys and talk about how sexy girls are and drool while discussing their body parts and wanting to “hit it”. A real man doesn’t treat women like a buffet. A real man pursues a WIFE, not a girlfriend and he goes about it the right way. If a Real man is interested in a female, he goes to their parents to let them know his intentions, like how they did it back in the days. A real man works on lowering his gaze when beautiful women walk by. A real man keeps his interactions with women short, cordial and to the point and doesn’t let it go longer so flirting and dirty thoughts can come into the mix. If a man is married, he stays faithful and doesn’t leer his eyes elsewhere – and that includes the TV, the internet and magazines. A real man is respectful to his wife, both in public and in private. A real man does not raise his hands to his wife no matter what the case is. A real man doesn’t point out his wife’s flaws and treat her like a second-class citizen. A real man treats every female (that isn’t his wife) he comes in contact with in the same way he treats his mother or sister.
Trait # 8: A Real Man Keeps His House in Order
Contrary to popular belief, a real man doesn’t live in a pigsty. There aren’t any pizza boxes on the counter, the sink isn’t filled with dishes and his underwear isn’t laying around. A real man is the master of his domain if he’s living on his own, because if not then he better move back in with his mom and get some lessons if he ever wants to land a wife. That’s correct! A real man DOES CHORES, even with his wife around. He does the dishes if they need cleaning, takes out the garbage, does his own laundry, irons his clothes, cooks from time to time. If you haven’t reached that level, you better work on it now. The Prophet Muhammad regularly did household work and did most of his own chores himself like fixing his shoes, doing his own laundry etc. Stop using manliness as an excuse and get your job done, clean your place up, get your documents in order and clean up yourself. |
1. Field of the Invention
The present invention relates to a liquid crystal display device and a pixel inspection method therefor, and more particularly to a liquid crystal display device and a pixel inspection method therefor that perform gray scale display using the combination of a plurality of subframes according to gray scale levels expressed by a plurality of bits.
2. Description of the Related Art
Heretofore, a subframe driving method is known as one of halftone display methods in liquid crystal display devices. In a subframe driving method which is one type of time base modulation methods, a predetermined period (one frame that is a unit for display of one image in the case of moving pictures, for example) is split into a plurality of subframes, and pixels are driven in a combination of subframes according to a gray scale to be displayed. The gray scale to be displayed is determined according to the ratio of a pixel drive period occupied in a predetermined period, and this ratio is specified by the combination of subframes.
In liquid crystal display devices according to this subframe driving method, one is known in which pixels are individually configured of a master latch, a slave latch, a liquid crystal display element, and first to third switching transistors, three transistors in total (see Japanese Patent Application National Publication (Laid-Open) No. 2001-523847, for example). In this pixel, one bit of a first data is applied to one input terminal of two input terminals of the master latch through the first switching transistor, a second data in the complementary relation with the first data is applied to the other input terminal through the second switching transistor, and when the pixel is selected by a row select signal applied through a row scanning line, the first data is written as the first and second switching transistors are turned to the ON-state. For example, when the first data has the logical value “1” and the second data has the logical value “0”, the pixel performs display.
After the data is written to all the pixels through the similar operations described above, the data written to the master latch are simultaneously read to the slave latch as the third switching transistors of all the pixels are turned to the ON-state in the subframe period, and the data latched to the slave latch are applied from the slave latch to the pixel electrode of the liquid crystal display element. The operations above are then repeated for the individual subframes, and desired gray scale display is performed with the combinations of all the subframes in a frame period.
Namely, in the liquid crystal display device according to the subframe driving method, all of the subframes in a frame period are preallocated to the same predetermined period or a different predetermined period. In the pixels, display is performed on all the subframes in the maximum gray scale display, display is not performed on all the subframes in the minimum gray scale display, and subframes for display are selected according to the gray scale for display in the case of the other gray scales. In the previously existing liquid crystal display device, inputted data is digital data expressing a gray scale, and the method is also a digital driving method in a two-stage latch configuration.
However, in the previously existing liquid crystal display device, since the two latches in the pixels are configured of static random access memories (SRAM), the number of transistors is increased and it is difficult to downsize the pixels.
Moreover, in the pixel above, generally, a silicon backplane including shift registers, for example, is prepared in large-scale semiconductor integrated circuit (LSI) processes. However, in probe inspection after a wafer is prepared, there is a problem that pixel inspection is not performed normally. This is because there is a possibility that the SRAM is rewritten due to electric charges accumulated on a column data line. Because when the pixel inspection is performed, data written to the SRAM is read out from the column data line after the data is input to the column data line and the input data is written to the SRAM.
In the description of Japanese Patent Application National Publication (Laid-Open) No. 2001-523847, a two-switch SRAM including two complementary bit lines is described. In contrast to this, here, the case of a one-switch SRAM configured of a single bit line and a single switch is considered.
For example, in the case of a full high definition (FHD) liquid crystal display device, the number of pixels lengthwise on the screen is 1,080 pixels, and the capacitance of the individual column data lines is about 1 pF. For example, an SRAM is configured of a switching transistor connected to a column data line at zero volt at low level and two inverters in which an input terminal of one inverter is connected to an output terminal of the other inverter. In these two inverters, suppose that the voltage of the input terminal of the one inverter connected to the switching transistor is at high level at a voltage of 3.3 V. In this case, when the switching transistor is turned on, the column data lines are charged at about 1 pF of electric charge capacitance described above from a P-channel MOS field effect transistor (in the following, referred to as a P-MOS transistor) configuring the other inverter whose output terminal is connected to the switching transistor.
At this time, since the driving force of the transistor configuring the other inverter is smaller than the driving force of the transistor configuring the one inverter, charging time is prolonged to cause incomplete charging, the voltage of the input terminal of the one inverter is below the turnover voltage, and the voltage (namely, data that has to be written to the SRAM) of the input terminal of the one inverter is rewritten. Thus, data on the SRAM is not enabled to be output to the column data line, and pixels are not accurately inspected.
The present invention is made on the viewpoints above, and it is an object to provide a liquid crystal display device and a pixel inspection method therefor that can downsize a pixel as compared with a pixel using two SRAMs in the pixel and can accurately inspect pixels. |
Networks for medical nutrition education--a review of the US experience and future prospects.
Nutrition education for physicians in the United State and Canada, remains an orphan discipline with no improvement or changes in the past decade. This status has led to the creation of regional nutrition centers and networks to develop nutrition education programs in health professional schools. The expanded network concept now covers geographic areas in the northeastern, southeastern, and middle United States, with plans to extend to the far West and to establish a national coordinating center or "network of networks." The barriers to progress in nutrition education must be overcome, and their conquest could form the basis for initiating new regional networks or a coordinating center. Strategies for nutrition networking are applicable to other disciplines and provide guidance for creating consortia to deal with the diminished need for medical and surgical subspecialties and play an important role in the training of generalist physicians. |
Q:
How to view current display resolution
How can I find out the current resolution the screen is running at, in OS X 10.10?
Under settings -> display (which is where I think it used to be), it shows the refresh frequency, but not the resolution. Using the monitor's controls, I can see it's running at 3840x2160, but how would one find this out from OS X?
A:
You go to apple menu -> about this mac, and there is a Displays tab with the information.
A:
You can get the resolution in Terminal using system_profiler by issuing the following command:
system_profiler SPDisplaysDataType | awk '/Resolution/{print $2, $3, $4}'
A:
You can see the current resolution in the Display system preferences.
On a Macbook .. Retina, If you have scaled selected for resolution
Hover your mouse over the current scaled selection and the resolution will be shown.
If you have default selected for resolution then the resolution is not shown.
( I cannot check my non retina desktop at the moment)
But either way you could run this Applescript/Objective - C script from your Applescript Menu or from Script Editor.
use framework "CoreGraphics"
use scripting additions
set sizes to item 1 of (current application's NSScreen's mainScreen's frame as list)
set theRez to ("width:" & width of |size| of sizes & " x height:" & height of |size| of sizes) as string
display dialog theRez with title "Main Screen Rez" buttons "OK"
|
1. Field of the Invention
The present invention relates to a semiconductor device including an oxide semiconductor film and a method for manufacturing the semiconductor device.
Note that the semiconductor device in this specification refers to all devices that can function by utilizing semiconductor characteristics, and electro-optic devices, semiconductor circuits, and electronic appliances are all semiconductor devices.
2. Description of the Related Art
In recent years, a market for mobile communications typified by mobile phones and the like has grown rapidly. To meet increasing requirements such as low power consumption, high integration degree, many functions, and high speed in semiconductor integration circuits that are mounted components, improvements in transistor characteristics are needed.
As one of indexes denoting transistor characteristics, an S-value is given. An S-value represents the amount of a change in gate voltage which is needed for changing a drain current by one digit under constant drain voltage. As the S-value becomes smaller, the controllability of the drain current is increased. Under the present circumstances, it is difficult to make the S-value less than or equal to 80 mV/dec because of miniaturization of transistors and the like.
Attention has also been drawn to a technique by which a transistor is manufactured using an amorphous oxide semiconductor material instead of a silicon-based semiconductor material and is applied to an electronic device or the like. For example, a technique for manufacturing a transistor whose channel layer is formed using an amorphous oxide semiconductor material containing indium (In), gallium (Ga), and zinc (Zn) and having an electron carrier concentration lower than 1018/cm3 is disclosed (see Patent Document 1). |
Electroshock weapons - such as stun guns and other similar devices that temporarily incapacitate a person by delivering a high-voltage, low-current electric shock - have helped law enforcement officers safely subdue dangerous or violent persons for years. The use of these weapons has been challenged, however, by claims that they may have contributed to more than 150 deaths in the United States since 2001. Now, researchers at the National Institute of Standards and Technology (NIST) are working toward a standard method for accurately assessing the electrical output of these devices, the results of which can be used in establishing baselines for future medical and safety studies.
Groups such as Amnesty International have called for guidelines for electroshock weapons that include “threshold exposures” (the minimum charges that would incapacitate different groups of people without putting them at risk for injury or death). One obstacle to the development of such guidelines is the fact that various reports regarding the output of electroshock weapons—the current and voltage they deliver—are inconsistent.
To address this problem, scientists in NIST’s Office of Law Enforcement Standards (OLES) have developed methods for calibrating the high-voltage and current measurement probes used by industry so that any inherent biases in the probes are minimized. By compensating for these probe effects, voltage and current readings were obtained that reflect the energies being dispensed by the weapons.
Next steps in the characterization program for electroshock weapons include implementing a second type of high-voltage measurement to verify the probe calibration system; further refining the uncertainty analysis for the overall measurement method to better define its accuracy and reliability; and, eventually, working with government agencies and the law enforcement community to standardize the method that will facilitate establishment of use guidelines. |
<?xml version="1.0" encoding="UTF-8"?>
<!-- ~ Copyright 2012 Red Hat, Inc. and/or its affiliates ~ ~ Licensed under the Apache License, Version
2.0 (the "License"); ~ you may not use this file except in compliance with
the License. ~ You may obtain a copy of the License at ~ ~ http://www.apache.org/licenses/LICENSE-2.0
~ ~ Unless required by applicable law or agreed to in writing, software ~
distributed under the License is distributed on an "AS IS" BASIS, ~ WITHOUT
WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. ~ See the
License for the specific language governing permissions and ~ limitations
under the License. -->
<persistence version="2.1"
xmlns="http://xmlns.jcp.org/xml/ns/persistence"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://xmlns.jcp.org/xml/ns/persistence http://xmlns.jcp.org/xml/ns/persistence/persistence_2_1.xsd">
<persistence-unit name="org.jbpm.domain" transaction-type="JTA">
<provider>org.hibernate.jpa.HibernatePersistenceProvider</provider>
<jta-data-source>java:jboss/datasources/ExampleDS</jta-data-source>
<mapping-file>META-INF/Taskorm.xml</mapping-file>
<mapping-file>META-INF/JBPMorm.xml</mapping-file>
<mapping-file>META-INF/Servicesorm.xml</mapping-file>
<mapping-file>META-INF/TaskAuditorm.xml</mapping-file>
<mapping-file>META-INF/Executor-orm.xml</mapping-file>
<mapping-file>META-INF/CaseMgmtorm.xml</mapping-file>
<!-- executor service -->
<class>org.jbpm.executor.entities.ErrorInfo</class>
<class>org.jbpm.executor.entities.RequestInfo</class>
<!-- task service -->
<class>org.jbpm.services.task.impl.model.AttachmentImpl</class>
<class>org.jbpm.services.task.impl.model.ContentImpl</class>
<class>org.jbpm.services.task.impl.model.BooleanExpressionImpl</class>
<class>org.jbpm.services.task.impl.model.CommentImpl</class>
<class>org.jbpm.services.task.impl.model.DeadlineImpl</class>
<class>org.jbpm.services.task.impl.model.CommentImpl</class>
<class>org.jbpm.services.task.impl.model.DeadlineImpl</class>
<class>org.jbpm.services.task.impl.model.DelegationImpl</class>
<class>org.jbpm.services.task.impl.model.EscalationImpl</class>
<class>org.jbpm.services.task.impl.model.GroupImpl</class>
<class>org.jbpm.services.task.impl.model.I18NTextImpl</class>
<class>org.jbpm.services.task.impl.model.NotificationImpl</class>
<class>org.jbpm.services.task.impl.model.EmailNotificationImpl</class>
<class>org.jbpm.services.task.impl.model.EmailNotificationHeaderImpl</class>
<class>org.jbpm.services.task.impl.model.PeopleAssignmentsImpl</class>
<class>org.jbpm.services.task.impl.model.ReassignmentImpl</class>
<class>org.jbpm.services.task.impl.model.TaskImpl</class>
<class>org.jbpm.services.task.impl.model.TaskDefImpl</class>
<class>org.jbpm.services.task.impl.model.TaskDataImpl</class>
<class>org.jbpm.services.task.impl.model.UserImpl</class>
<class>org.jbpm.services.task.impl.model.EmailImpl</class>
<!-- Event Classes -->
<class>org.jbpm.services.task.audit.impl.model.TaskEventImpl</class>
<!-- Task Audit Classes -->
<class>org.jbpm.services.task.audit.impl.model.AuditTaskImpl</class>
<class>org.jbpm.services.task.audit.impl.model.TaskVariableImpl</class>
<!--BAM for task service -->
<class>org.jbpm.services.task.audit.impl.model.BAMTaskSummaryImpl</class>
<!-- engine -->
<class>org.drools.persistence.info.SessionInfo</class>
<class>org.jbpm.persistence.processinstance.ProcessInstanceInfo</class>
<class>org.drools.persistence.info.WorkItemInfo</class>
<class>org.jbpm.persistence.correlation.CorrelationKeyInfo</class>
<class>org.jbpm.persistence.correlation.CorrelationPropertyInfo</class>
<!-- manager -->
<class>org.jbpm.runtime.manager.impl.jpa.ContextMappingInfo</class>
<!-- bam -->
<class>org.jbpm.process.audit.ProcessInstanceLog</class>
<class>org.jbpm.process.audit.NodeInstanceLog</class>
<class>org.jbpm.process.audit.VariableInstanceLog</class>
<!-- deployment store -->
<class>org.jbpm.kie.services.impl.store.DeploymentStoreEntry</class>
<!-- query service storage -->
<class>org.jbpm.kie.services.impl.query.persistence.QueryDefinitionEntity</class>
<!-- case id generation -->
<class>org.jbpm.casemgmt.impl.generator.CaseIdInfo</class>
<class>org.jbpm.casemgmt.impl.audit.CaseRoleAssignmentLog</class>
<class>org.jbpm.casemgmt.impl.audit.CaseFileDataLog</class>
<!-- error handling -->
<class>org.jbpm.runtime.manager.impl.jpa.ExecutionErrorInfo</class>
<properties>
<property name="hibernate.dialect" value="org.hibernate.dialect.H2Dialect" />
<property name="hibernate.max_fetch_depth" value="3" />
<property name="hibernate.hbm2ddl.auto" value="update" />
<property name="hibernate.show_sql" value="false" />
<!-- BZ 841786: AS7/EAP 6/Hib 4 uses new (sequence) generators which seem to cause problems -->
<property name="hibernate.id.new_generator_mappings" value="false" />
<property name="hibernate.transaction.jta.platform" value="JBossAS" />
</properties>
</persistence-unit>
</persistence>
|
Effectiveness of case management for community elderly with hypertension, diabetes mellitus, and hypercholesterolemia in Taiwan: a record review.
To identify the effectiveness of case management for community elderly with hypertension (HT), diabetes mellitus, and hypercholesterolemia (HC) (the so-called three highs). Secondary data of the first and 3-month-after visiting records were extracted from 33 Public Health Centers in Taiwan. Seven hundred and sixty-six clients were selected who were at least 65 years old and had been diagnosed twice on the Case Management Record with at least one of the three highs. This sample had a mean age of 72.6 years, 59.7% were female. Approximately 74% of the clients had HT, 55% had diabetes, and 15% had HC. Each elderly revealed 1.4 highs of the three highs. The elderly with HT, and diabetes, their blood pressures (BP) and blood sugars significantly decreased after being managed by public health nurses. Males and the elderly living in urban areas had more decrease in systolic BP. Females had more decrease in fasting blood sugar. The study found that the case management of the three highs presented effectiveness on reducing the values of the three highs of the elderly in community. |
Contact Us
Carberry
Carberry aims to be a shield for Raspberry Pi microcomputers. Carberry represents the link between car electronics and Raspberry Pi, which allows the development of end-user applications, such as media centers, vehicle diagnostics, data logging, fleet management, tracking, blackboxes, burglar alarms, carputing, internet, and much more.
Raspberry Pi B Version
Product Code:
Raspberry Pi 2/B+ Version
Product Code: CF0066UNCY31
Projects: give power to ideas!
CarPC
CarPC based on Raspberry are now reality, you could check an example here.
Carberry integration allows you to add steering wheel commands functionality and also other information like: Positive Ignition, Lights, Speedpulse... information.
GPS Monitoring System
Connecting Carberry to car OBD port and using a GPS antenna connected to Raspberry Pi is possible to create a vehicle tracking system. Speed, Fuel Level, Engine Load, etc... could be read from OBD port and saved insiede carberry memory, combined with GPS position. Adding a 3G modem is possible to create a live tracking system.
Hardware necessary:
Carberry
Raspberry Pi
GPS antenna
3G modem
Blackbox
Carberry has onboard an accelerometer + magnetometer sensor that allows you to record G axis movement related to car speed, in this way is possible to create a blackbox to track car behavior during a crash.
Hardware necessary:
Carberry
Raspberry Pi
GPS antenna
3G modem
Car Diagnostic
Connecting Carberry to car OBD port you can get real time data (speed, RPM, fuel level, turbo pressure...) and diagnostic errors.
This allows to create a car diagnostic computer or a live data info panel like in our project.
Please check our project: Carberry Info Panel
Hardware necessary:
Carberry
Raspberry Pi
Video in motion
Carberry has 2 Canbus and 2 GMLAN lines: you can use one as input and one as output filtering or altering which information you want from the input to ouput.
Using this technique you could alter speedpulse information in order to unlock video in motion on original infotainment systems while driving.
Hardware necessary:
Carberry
Raspberry Pi
Alarm System
Using Carberry onboard accelerometer + magnetometer sensors is possible to recognizes car movement when is closed, comining these information with GPS position is possible make a full car alarm system.
Hardware necessary:
Carberry
Raspberry Pi
GPS antenna
3G modem
G Meter
Carberry accelerometer could be used to create a powerful G Meter and Raspberry pi could draw a the graphic to any kind of display.
Hardware necessary:
Carberry
Raspberry Pi
Tyre Pressure Monitoring
In different cars is available speed info of each wheel, calculating different rotating speed is possible to understand if a one tyre is deflated than others and generate an alert for the driver.
A similar product, created with dedicated hardware, was already developed by Paser.
Please check product page: SAFETY KIT SKT170
Hardware necessary:
Carberry
Raspberry Pi
4x50W Amplifier
1x Speaker
Carberry uses are not just those examples, real limit is just your imagination...
How It Works
Carberry is connected to Raspberry Pi via the expansion port P1, by this connection Carberry is able to supply Rasberry Pi.
The communication between Raspberry Pi operating system of and Carberry takes place via the serial port.
By default it is provided a daemon that allows communication via a TCP/IP socket that responds on port 7070, you can anyway develop your own custom daemon to communicate with Carberry directly on the serial port.
Having a dedicated approach to the cars, Carberry manages Raspberry Pi supply, via a power remote control (whether it is analog on dedicated wire, on the CANBUS or line GMLAN).
By acting on the power remote control Carberry will feed the Rasberry Pi or if necessary will launch the shutdown of the operating system.
Communication stack between Carberry and Raspberry Pi is the following:
Board Specification
CPU
32 bit RISC Microcontroller
Supply
12V Car supply
Consumption 150mA / less than 3ma in standby
BUSes
2xCAN BUS
2x GMLAN
2x LADDER
GPIO
2xGPI
2xGPO
1xGPIO CMOS
1xIGNITION IN
1xIGNITION OUT
User Interface
1xReset button
1xBicolor status led
2xGeneral Purpose LED
UART
One TTL 5V RX/TX UART
USB device (Virtual COM Port)
Storage
EEPROM
SRAM
Sensors
Accelerometer
Magnetometer
RTCC
Multimedia
40KHz Infrared emitter
40KHz Infrared receiver
MFI Apple Chip Support
Dimensions
Raspberry Pi shaped
Weight
Hardware specifications
Same size printed circuit board as of Raspberry Pi, with proper shaping.
Connection to Raspberry via the P1 26 pin GPIO header located on the board.
22 pin Microfit connector for vehicle connection.
Power supply +12 V and GND from the vehicle.
Power supply 5V 1A for Rasperry Pi generated onboard by Carberry.
Control of the Raspberry Pi power supply by mosfet, for the management of low consumption.
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
A1
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
A3
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
A3
Year: 2004-2006
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
A3
Year: 2007-2012
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
A4
Year: 2004-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
A4
Year: 2008-2013
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
A5
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
A6
Year: 2006-2013
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
A8
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Q5
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Q7
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
TT
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Series 1
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Series 1
Year: 2005-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Series 3
Year: 2005-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Series 5
Year: 2005-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Series 5
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Series 5 GT
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
X1
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
X3
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
X5
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
X6
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
BLS
Year: 2007-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Escalade
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Aveo
Year: 2008-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Captiva
Year: 2006-2012
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Captiva
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Cruze
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Epica
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Evanda
Year: 2003-2005
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Kalos
Year: 2002-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Matiz
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Tacuma
Year: 2004-2008
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Tahoe
Year: 2006-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
300C
Year: 2005-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
300C
Year: 2008-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
300C
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Grand Voyager
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
C Crosser
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
C2
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
C3
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
C3 Picasso
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
DS3
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Terios
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Caliber
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Journey
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Nitro
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Ram
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
California
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
16
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
500
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
500L
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Barchetta
Year: 2003-2005
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Bravo
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Ducato
Year: 2007-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Ducato
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Freemont
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Grande Punto
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Idea
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Marea
Year: 2003-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Multipla
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
New Croma
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Panda
Year: 2004-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Panda
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Punto Evo
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Qubo
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Scudo
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Scudo
Year: 2004-2004
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Stilo
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
B-Max
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
C-Max
Year: 2006-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
C-Max
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Fiesta
Year: 2002-2005
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Fiesta
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Fiesta
Year: 2006-2008
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Focus
Year: 2002-2003
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Focus
Year: 2004-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Focus
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Kuga
Year: 2008-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Mondeo
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
S-Max
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Tourneo
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Transit
Year: 2007-2012
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Accord
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Civic
Year: 2006-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Civic
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Cr-V
Year: 2006-2012
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Cr-V
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Fr-V
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Santa Fe
Year: 2006-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Santa Fe
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Veloster
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
i20
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
i30
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
i40
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
ix20
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
ix35
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
ix55
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Dmax
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Stralis
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
S-Type
Year: 2004-2008
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
X-Type
Year: 2004-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
XF
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Cherokee
Year: 2001-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Cherokee
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Compass
Year: 2006-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Compass
Year: 2010-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Compass
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Grand Cherokee
Year: 2005-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Grand Cherokee
Year: 2003-2005
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Grand Cherokee
Year: 2010-2013
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Grand Cherokee
Year: 2008-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Wrangler
Year: 2007-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Wrangler
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Carens
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Cee'd
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Cee'd
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Sorento
Year: 2006-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Sorento
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Soul
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Sportage
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Venga
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Delta
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Musa
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Ypsilon
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Freelander 2
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Range Rover Sport
Year: 2008-2013
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Range Rover Vogue
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Cx7
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Mazda 2
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Mazda 3
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Mazda 5
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Mazda 6
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Actros
Year: 2007-2013
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class A
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class A Fakra
Year: 2003-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class A ISO
Year: 2003-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class B
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class C
Year: 2003-2014
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class CLK
Year: 2003-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class E
Year: 2004-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class E Coupe
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class GL
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class ML
Year: 2004-2005
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class R
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class S
Year: 2006-2008
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class SL
Year: 2006-2008
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Class SLK
Year: 2003-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
New Ml
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Sprinter Fakra
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Sprinter ISO
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Viano
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Vito
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Mini
Year: 2001-2006
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Mini
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Mini Clubman
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Mini Countryman
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
ASX
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
L200
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Lancer
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Outlander
Year: 2007-2012
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Pajero
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
350z
Year: 2003-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
370z
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Evalia
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Juke
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Micra
Year: 2007-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Murano
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Navara
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Note
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Note
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Pathfinder
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Qashqai
Year: 2007-2013
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
X-Trail
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Agila
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
All Group
Year: 2001-2004
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Antara
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Astra H
Year: 2004-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Astra J
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Corsa
Year: 2004-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Corsa D
Year: 2008-2014
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Insignia
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Meriva
Year: 2004-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Meriva
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Signum
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Tigra
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Vectra
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Vivaro
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Zafira
Year: 2004-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Zafira Tourer
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
207
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
208
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
307
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
308
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
407
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
508
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
3008
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
4007
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
5008
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Boxter
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Cayenne
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Cayenne Turbo
Year: 2006-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Clio
Year: 2006-2008
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Clio
Year: 2009-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Koleos
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Laguna
Year: 2006-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Laguna III
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Megane
Year: 2006-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Megane III
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Modus
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Scenic Xmod
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Traffic
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Twingo
Year: 2009-2012
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Twingo
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
9.3
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
9.5
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
9.5
Year: 2006-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
R620
Year: 2005-2008
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Alhambra
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Altea
Year: 2004-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Altea
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Cordoba
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Ibiza
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Leon
Year: 2005-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Leon
Year: 2010-2012
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Leon
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Octavia II
Year: 2006-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Octavia II
Year: 2010-2012
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Octavia III
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Roomster
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Superb
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Yeti
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Actyon
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Korando
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Kyron
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Rexton II
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Forester
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Impreza
Year: 2006-2008
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Impreza
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Legacy
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Outback
Year: 2010-2011
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Trezia
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
XV
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Grand Vitara
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Splash
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Swift
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Sx4
Year: 2006-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Auris
Year: 2007-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Avensis
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Corolla
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Hilux
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Rav 4
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Verso
Year: 2009-2012
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Verso
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Yaris
Year: 2005-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
Amarok
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Beetle
Year: 2012-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Eos
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Golf V
Year: 2004-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Golf V Plus
Year: 2004-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Golf VI
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Golf VII
Year: 2013-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Multivan
Year: 2007-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
New Touareg
Year: 2007-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Passat
Year: 2004-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Passat
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Passat CC
Year: 2008-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Polo
Year: 2009-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Scirocco
Year: 2008-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Scirocco
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Sharan
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Tiguan
Year: 2007-2010
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Tiguan
Year: 2011-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Touareg
Year: 2004-2006
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Touran
Year: 2004-2009
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Touran
Year: 2010-2019
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
V70
Year: 2004-2007
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
V70
Year: 2002-2004
SW Commands
Positive Ignition
Illumination Signal
Speedpulse Signal
Reverse Gear
Negative Handbrake
Phone Button
Although Paser tries to keep the information on its website constantly updated, no guarantees are given about the website accuracy. Paser disclaims all responsibility for any error or omission into this website contents.
We inform you that Paser will be closed for Summer holidays from 29th July 2019 to 25th August 2019 included.
We invite you to send us your orders as soon as possible in order to give you the best service! Otherwise, we can\'t guarantee the deliveries before summer break and we will have to postpone them.
Thanks in advance for your collaboration and good holidays from all Paser Staff! |
Clonidine effect in chronic angina pectoris. Double-blind, crossover trial on 60 patients.
Increased adrenergic activity, often manifested in chronic angina, is likely to influence adversely the course of the disease. In view of the inhibitory effect of clonidine (CL) upon the adrenergic nervous system, the effectiveness of small doses of CL in chronic angina was evaluated in a double-blind crossover study on 60 patients suffering at least 5 coronary pains per week in spite of routine medication. CL was given orally in a dose of 2 x 75 microgram/day for a 2 wk. Reduction in frequency of coronary pains by at least 50% was observed in 53.7% of patients, total nitroglycerin consumption decreased from 322 to 174 tablets/week, and ergometric performance increased from 168 to 283 W x min/patient. Urinary excretion of adrenaline and noradrenaline diminished. Blood pressure and heart rate were not considerably changed. Mild and transient side effects occurred in 10 patients, 9 of them completed the trial. It is concluded that CL in low doses is effective and safe in patients with chronic angina, presumably by alleviating adrenergic strain. |
/*
* RISC-V CPU helpers for qemu.
*
* Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
* Copyright (c) 2017-2018 SiFive, Inc.
*
* This program is free software; you can redistribute it and/or modify it
* under the terms and conditions of the GNU General Public License,
* version 2 or later, as published by the Free Software Foundation.
*
* This program is distributed in the hope it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
* more details.
*
* You should have received a copy of the GNU General Public License along with
* this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "qemu/osdep.h"
#include "qemu/log.h"
#include "cpu.h"
#include "exec/exec-all.h"
#include "tcg-op.h"
#include "trace.h"
int riscv_cpu_mmu_index(CPURISCVState *env, bool ifetch)
{
#ifdef CONFIG_USER_ONLY
return 0;
#else
return env->priv;
#endif
}
#ifndef CONFIG_USER_ONLY
static int riscv_cpu_local_irq_pending(CPURISCVState *env)
{
target_ulong mstatus_mie = get_field(env->mstatus, MSTATUS_MIE);
target_ulong mstatus_sie = get_field(env->mstatus, MSTATUS_SIE);
target_ulong pending = atomic_read(&env->mip) & env->mie;
target_ulong mie = env->priv < PRV_M || (env->priv == PRV_M && mstatus_mie);
target_ulong sie = env->priv < PRV_S || (env->priv == PRV_S && mstatus_sie);
target_ulong irqs = (pending & ~env->mideleg & -mie) |
(pending & env->mideleg & -sie);
if (irqs) {
return ctz64(irqs); /* since non-zero */
} else {
return EXCP_NONE; /* indicates no pending interrupt */
}
}
#endif
bool riscv_cpu_exec_interrupt(CPUState *cs, int interrupt_request)
{
#if !defined(CONFIG_USER_ONLY)
if (interrupt_request & CPU_INTERRUPT_HARD) {
RISCVCPU *cpu = RISCV_CPU(cs);
CPURISCVState *env = &cpu->env;
int interruptno = riscv_cpu_local_irq_pending(env);
if (interruptno >= 0) {
cs->exception_index = RISCV_EXCP_INT_FLAG | interruptno;
riscv_cpu_do_interrupt(cs);
return true;
}
}
#endif
return false;
}
#if !defined(CONFIG_USER_ONLY)
int riscv_cpu_claim_interrupts(RISCVCPU *cpu, uint32_t interrupts)
{
CPURISCVState *env = &cpu->env;
if (env->miclaim & interrupts) {
return -1;
} else {
env->miclaim |= interrupts;
return 0;
}
}
/* iothread_mutex must be held */
uint32_t riscv_cpu_update_mip(RISCVCPU *cpu, uint32_t mask, uint32_t value)
{
CPURISCVState *env = &cpu->env;
uint32_t old, new, cmp = atomic_read(&env->mip);
do {
old = cmp;
new = (old & ~mask) | (value & mask);
cmp = atomic_cmpxchg(&env->mip, old, new);
} while (old != cmp);
if (new) {
cpu_interrupt(CPU(cpu), CPU_INTERRUPT_HARD);
} else {
cpu_reset_interrupt(CPU(cpu), CPU_INTERRUPT_HARD);
}
return old;
}
void riscv_cpu_set_mode(CPURISCVState *env, target_ulong newpriv)
{
if (newpriv > PRV_M) {
g_assert_not_reached();
}
if (newpriv == PRV_H) {
newpriv = PRV_U;
}
/* tlb_flush is unnecessary as mode is contained in mmu_idx */
env->priv = newpriv;
}
/* get_physical_address - get the physical address for this virtual address
*
* Do a page table walk to obtain the physical address corresponding to a
* virtual address. Returns 0 if the translation was successful
*
* Adapted from Spike's mmu_t::translate and mmu_t::walk
*
*/
static int get_physical_address(CPURISCVState *env, hwaddr *physical,
int *prot, target_ulong addr,
int access_type, int mmu_idx)
{
/* NOTE: the env->pc value visible here will not be
* correct, but the value visible to the exception handler
* (riscv_cpu_do_interrupt) is correct */
int mode = mmu_idx;
if (mode == PRV_M && access_type != MMU_INST_FETCH) {
if (get_field(env->mstatus, MSTATUS_MPRV)) {
mode = get_field(env->mstatus, MSTATUS_MPP);
}
}
if (mode == PRV_M || !riscv_feature(env, RISCV_FEATURE_MMU)) {
*physical = addr;
*prot = PAGE_READ | PAGE_WRITE | PAGE_EXEC;
return TRANSLATE_SUCCESS;
}
*prot = 0;
target_ulong base;
int levels, ptidxbits, ptesize, vm, sum;
int mxr = get_field(env->mstatus, MSTATUS_MXR);
if (env->priv_ver >= PRIV_VERSION_1_10_0) {
base = get_field(env->satp, SATP_PPN) << PGSHIFT;
sum = get_field(env->mstatus, MSTATUS_SUM);
vm = get_field(env->satp, SATP_MODE);
switch (vm) {
case VM_1_10_SV32:
levels = 2; ptidxbits = 10; ptesize = 4; break;
case VM_1_10_SV39:
levels = 3; ptidxbits = 9; ptesize = 8; break;
case VM_1_10_SV48:
levels = 4; ptidxbits = 9; ptesize = 8; break;
case VM_1_10_SV57:
levels = 5; ptidxbits = 9; ptesize = 8; break;
case VM_1_10_MBARE:
*physical = addr;
*prot = PAGE_READ | PAGE_WRITE | PAGE_EXEC;
return TRANSLATE_SUCCESS;
default:
g_assert_not_reached();
}
} else {
base = env->sptbr << PGSHIFT;
sum = !get_field(env->mstatus, MSTATUS_PUM);
vm = get_field(env->mstatus, MSTATUS_VM);
switch (vm) {
case VM_1_09_SV32:
levels = 2; ptidxbits = 10; ptesize = 4; break;
case VM_1_09_SV39:
levels = 3; ptidxbits = 9; ptesize = 8; break;
case VM_1_09_SV48:
levels = 4; ptidxbits = 9; ptesize = 8; break;
case VM_1_09_MBARE:
*physical = addr;
*prot = PAGE_READ | PAGE_WRITE | PAGE_EXEC;
return TRANSLATE_SUCCESS;
default:
g_assert_not_reached();
}
}
CPUState *cs = CPU(riscv_env_get_cpu(env));
int va_bits = PGSHIFT + levels * ptidxbits;
target_ulong mask = (1L << (TARGET_LONG_BITS - (va_bits - 1))) - 1;
target_ulong masked_msbs = (addr >> (va_bits - 1)) & mask;
if (masked_msbs != 0 && masked_msbs != mask) {
return TRANSLATE_FAIL;
}
int ptshift = (levels - 1) * ptidxbits;
int i;
#if !TCG_OVERSIZED_GUEST
restart:
#endif
for (i = 0; i < levels; i++, ptshift -= ptidxbits) {
target_ulong idx = (addr >> (PGSHIFT + ptshift)) &
((1 << ptidxbits) - 1);
/* check that physical address of PTE is legal */
target_ulong pte_addr = base + idx * ptesize;
#if defined(TARGET_RISCV32)
target_ulong pte = ldl_phys(cs->as, pte_addr);
#elif defined(TARGET_RISCV64)
target_ulong pte = ldq_phys(cs->as, pte_addr);
#endif
target_ulong ppn = pte >> PTE_PPN_SHIFT;
if (!(pte & PTE_V)) {
/* Invalid PTE */
return TRANSLATE_FAIL;
} else if (!(pte & (PTE_R | PTE_W | PTE_X))) {
/* Inner PTE, continue walking */
base = ppn << PGSHIFT;
} else if ((pte & (PTE_R | PTE_W | PTE_X)) == PTE_W) {
/* Reserved leaf PTE flags: PTE_W */
return TRANSLATE_FAIL;
} else if ((pte & (PTE_R | PTE_W | PTE_X)) == (PTE_W | PTE_X)) {
/* Reserved leaf PTE flags: PTE_W + PTE_X */
return TRANSLATE_FAIL;
} else if ((pte & PTE_U) && ((mode != PRV_U) &&
(!sum || access_type == MMU_INST_FETCH))) {
/* User PTE flags when not U mode and mstatus.SUM is not set,
or the access type is an instruction fetch */
return TRANSLATE_FAIL;
} else if (!(pte & PTE_U) && (mode != PRV_S)) {
/* Supervisor PTE flags when not S mode */
return TRANSLATE_FAIL;
} else if (ppn & ((1ULL << ptshift) - 1)) {
/* Misaligned PPN */
return TRANSLATE_FAIL;
} else if (access_type == MMU_DATA_LOAD && !((pte & PTE_R) ||
((pte & PTE_X) && mxr))) {
/* Read access check failed */
return TRANSLATE_FAIL;
} else if (access_type == MMU_DATA_STORE && !(pte & PTE_W)) {
/* Write access check failed */
return TRANSLATE_FAIL;
} else if (access_type == MMU_INST_FETCH && !(pte & PTE_X)) {
/* Fetch access check failed */
return TRANSLATE_FAIL;
} else {
/* if necessary, set accessed and dirty bits. */
target_ulong updated_pte = pte | PTE_A |
(access_type == MMU_DATA_STORE ? PTE_D : 0);
/* Page table updates need to be atomic with MTTCG enabled */
if (updated_pte != pte) {
/*
* - if accessed or dirty bits need updating, and the PTE is
* in RAM, then we do so atomically with a compare and swap.
* - if the PTE is in IO space or ROM, then it can't be updated
* and we return TRANSLATE_FAIL.
* - if the PTE changed by the time we went to update it, then
* it is no longer valid and we must re-walk the page table.
*/
MemoryRegion *mr;
hwaddr l = sizeof(target_ulong), addr1;
mr = address_space_translate(cs->as, pte_addr,
&addr1, &l, false, MEMTXATTRS_UNSPECIFIED);
if (memory_region_is_ram(mr)) {
target_ulong *pte_pa =
qemu_map_ram_ptr(mr->ram_block, addr1);
#if TCG_OVERSIZED_GUEST
/* MTTCG is not enabled on oversized TCG guests so
* page table updates do not need to be atomic */
*pte_pa = pte = updated_pte;
#else
target_ulong old_pte =
atomic_cmpxchg(pte_pa, pte, updated_pte);
if (old_pte != pte) {
goto restart;
} else {
pte = updated_pte;
}
#endif
} else {
/* misconfigured PTE in ROM (AD bits are not preset) or
* PTE is in IO space and can't be updated atomically */
return TRANSLATE_FAIL;
}
}
/* for superpage mappings, make a fake leaf PTE for the TLB's
benefit. */
target_ulong vpn = addr >> PGSHIFT;
*physical = (ppn | (vpn & ((1L << ptshift) - 1))) << PGSHIFT;
/* set permissions on the TLB entry */
if ((pte & PTE_R) || ((pte & PTE_X) && mxr)) {
*prot |= PAGE_READ;
}
if ((pte & PTE_X)) {
*prot |= PAGE_EXEC;
}
/* add write permission on stores or if the page is already dirty,
so that we TLB miss on later writes to update the dirty bit */
if ((pte & PTE_W) &&
(access_type == MMU_DATA_STORE || (pte & PTE_D))) {
*prot |= PAGE_WRITE;
}
return TRANSLATE_SUCCESS;
}
}
return TRANSLATE_FAIL;
}
static void raise_mmu_exception(CPURISCVState *env, target_ulong address,
MMUAccessType access_type)
{
CPUState *cs = CPU(riscv_env_get_cpu(env));
int page_fault_exceptions =
(env->priv_ver >= PRIV_VERSION_1_10_0) &&
get_field(env->satp, SATP_MODE) != VM_1_10_MBARE;
switch (access_type) {
case MMU_INST_FETCH:
cs->exception_index = page_fault_exceptions ?
RISCV_EXCP_INST_PAGE_FAULT : RISCV_EXCP_INST_ACCESS_FAULT;
break;
case MMU_DATA_LOAD:
cs->exception_index = page_fault_exceptions ?
RISCV_EXCP_LOAD_PAGE_FAULT : RISCV_EXCP_LOAD_ACCESS_FAULT;
break;
case MMU_DATA_STORE:
cs->exception_index = page_fault_exceptions ?
RISCV_EXCP_STORE_PAGE_FAULT : RISCV_EXCP_STORE_AMO_ACCESS_FAULT;
break;
default:
g_assert_not_reached();
}
env->badaddr = address;
}
hwaddr riscv_cpu_get_phys_page_debug(CPUState *cs, vaddr addr)
{
RISCVCPU *cpu = RISCV_CPU(cs);
hwaddr phys_addr;
int prot;
int mmu_idx = cpu_mmu_index(&cpu->env, false);
if (get_physical_address(&cpu->env, &phys_addr, &prot, addr, 0, mmu_idx)) {
return -1;
}
return phys_addr;
}
void riscv_cpu_do_unaligned_access(CPUState *cs, vaddr addr,
MMUAccessType access_type, int mmu_idx,
uintptr_t retaddr)
{
RISCVCPU *cpu = RISCV_CPU(cs);
CPURISCVState *env = &cpu->env;
switch (access_type) {
case MMU_INST_FETCH:
cs->exception_index = RISCV_EXCP_INST_ADDR_MIS;
break;
case MMU_DATA_LOAD:
cs->exception_index = RISCV_EXCP_LOAD_ADDR_MIS;
break;
case MMU_DATA_STORE:
cs->exception_index = RISCV_EXCP_STORE_AMO_ADDR_MIS;
break;
default:
g_assert_not_reached();
}
env->badaddr = addr;
riscv_raise_exception(env, cs->exception_index, retaddr);
}
/* called by qemu's softmmu to fill the qemu tlb */
void tlb_fill(CPUState *cs, target_ulong addr, int size,
MMUAccessType access_type, int mmu_idx, uintptr_t retaddr)
{
int ret;
ret = riscv_cpu_handle_mmu_fault(cs, addr, size, access_type, mmu_idx);
if (ret == TRANSLATE_FAIL) {
RISCVCPU *cpu = RISCV_CPU(cs);
CPURISCVState *env = &cpu->env;
riscv_raise_exception(env, cs->exception_index, retaddr);
}
}
#endif
int riscv_cpu_handle_mmu_fault(CPUState *cs, vaddr address, int size,
int rw, int mmu_idx)
{
RISCVCPU *cpu = RISCV_CPU(cs);
CPURISCVState *env = &cpu->env;
#if !defined(CONFIG_USER_ONLY)
hwaddr pa = 0;
int prot;
#endif
int ret = TRANSLATE_FAIL;
qemu_log_mask(CPU_LOG_MMU,
"%s pc " TARGET_FMT_lx " ad %" VADDR_PRIx " rw %d mmu_idx \
%d\n", __func__, env->pc, address, rw, mmu_idx);
#if !defined(CONFIG_USER_ONLY)
ret = get_physical_address(env, &pa, &prot, address, rw, mmu_idx);
qemu_log_mask(CPU_LOG_MMU,
"%s address=%" VADDR_PRIx " ret %d physical " TARGET_FMT_plx
" prot %d\n", __func__, address, ret, pa, prot);
if (riscv_feature(env, RISCV_FEATURE_PMP) &&
!pmp_hart_has_privs(env, pa, TARGET_PAGE_SIZE, 1 << rw)) {
ret = TRANSLATE_FAIL;
}
if (ret == TRANSLATE_SUCCESS) {
tlb_set_page(cs, address & TARGET_PAGE_MASK, pa & TARGET_PAGE_MASK,
prot, mmu_idx, TARGET_PAGE_SIZE);
} else if (ret == TRANSLATE_FAIL) {
raise_mmu_exception(env, address, rw);
}
#else
switch (rw) {
case MMU_INST_FETCH:
cs->exception_index = RISCV_EXCP_INST_PAGE_FAULT;
break;
case MMU_DATA_LOAD:
cs->exception_index = RISCV_EXCP_LOAD_PAGE_FAULT;
break;
case MMU_DATA_STORE:
cs->exception_index = RISCV_EXCP_STORE_PAGE_FAULT;
break;
}
#endif
return ret;
}
/*
* Handle Traps
*
* Adapted from Spike's processor_t::take_trap.
*
*/
void riscv_cpu_do_interrupt(CPUState *cs)
{
#if !defined(CONFIG_USER_ONLY)
RISCVCPU *cpu = RISCV_CPU(cs);
CPURISCVState *env = &cpu->env;
/* cs->exception is 32-bits wide unlike mcause which is XLEN-bits wide
* so we mask off the MSB and separate into trap type and cause.
*/
bool async = !!(cs->exception_index & RISCV_EXCP_INT_FLAG);
target_ulong cause = cs->exception_index & RISCV_EXCP_INT_MASK;
target_ulong deleg = async ? env->mideleg : env->medeleg;
target_ulong tval = 0;
static const int ecall_cause_map[] = {
[PRV_U] = RISCV_EXCP_U_ECALL,
[PRV_S] = RISCV_EXCP_S_ECALL,
[PRV_H] = RISCV_EXCP_H_ECALL,
[PRV_M] = RISCV_EXCP_M_ECALL
};
if (!async) {
/* set tval to badaddr for traps with address information */
switch (cause) {
case RISCV_EXCP_INST_ADDR_MIS:
case RISCV_EXCP_INST_ACCESS_FAULT:
case RISCV_EXCP_LOAD_ADDR_MIS:
case RISCV_EXCP_STORE_AMO_ADDR_MIS:
case RISCV_EXCP_LOAD_ACCESS_FAULT:
case RISCV_EXCP_STORE_AMO_ACCESS_FAULT:
case RISCV_EXCP_INST_PAGE_FAULT:
case RISCV_EXCP_LOAD_PAGE_FAULT:
case RISCV_EXCP_STORE_PAGE_FAULT:
tval = env->badaddr;
break;
default:
break;
}
/* ecall is dispatched as one cause so translate based on mode */
if (cause == RISCV_EXCP_U_ECALL) {
assert(env->priv <= 3);
cause = ecall_cause_map[env->priv];
}
}
trace_riscv_trap(env->mhartid, async, cause, env->pc, tval, cause < 16 ?
(async ? riscv_intr_names : riscv_excp_names)[cause] : "(unknown)");
if (env->priv <= PRV_S &&
cause < TARGET_LONG_BITS && ((deleg >> cause) & 1)) {
/* handle the trap in S-mode */
target_ulong s = env->mstatus;
s = set_field(s, MSTATUS_SPIE, env->priv_ver >= PRIV_VERSION_1_10_0 ?
get_field(s, MSTATUS_SIE) : get_field(s, MSTATUS_UIE << env->priv));
s = set_field(s, MSTATUS_SPP, env->priv);
s = set_field(s, MSTATUS_SIE, 0);
env->mstatus = s;
env->scause = cause | ~(((target_ulong)-1) >> async);
env->sepc = env->pc;
env->sbadaddr = tval;
env->pc = (env->stvec >> 2 << 2) +
((async && (env->stvec & 3) == 1) ? cause * 4 : 0);
riscv_cpu_set_mode(env, PRV_S);
} else {
/* handle the trap in M-mode */
target_ulong s = env->mstatus;
s = set_field(s, MSTATUS_MPIE, env->priv_ver >= PRIV_VERSION_1_10_0 ?
get_field(s, MSTATUS_MIE) : get_field(s, MSTATUS_UIE << env->priv));
s = set_field(s, MSTATUS_MPP, env->priv);
s = set_field(s, MSTATUS_MIE, 0);
env->mstatus = s;
env->mcause = cause | ~(((target_ulong)-1) >> async);
env->mepc = env->pc;
env->mbadaddr = tval;
env->pc = (env->mtvec >> 2 << 2) +
((async && (env->mtvec & 3) == 1) ? cause * 4 : 0);
riscv_cpu_set_mode(env, PRV_M);
}
/* NOTE: it is not necessary to yield load reservations here. It is only
* necessary for an SC from "another hart" to cause a load reservation
* to be yielded. Refer to the memory consistency model section of the
* RISC-V ISA Specification.
*/
#endif
cs->exception_index = EXCP_NONE; /* mark handled to qemu */
}
|
Google’s social network, called Google+, is growing in terms of users and interaction, serving as the search giant’s competitor against its more established rivals in Facebook and Twitter.
Whilst the service draws many parallels with Facebook, one of Google+’s standout features is Hangouts; a video chat room that allows any Google user to connect from YouTube, Gmail, iGoogle and Orkut (and other Google properties) and chat to their friends, hold meetings or meet people for the first time using their webcam and a microphone.
The service supports up to 10 people chatting at any one time.
Many believe Google introduced the feature to compete with Skype, the popular app-based VOIP voice calling software, using the hook that it did not require additional clients and could be initiated in both the browser and via the official Google+ mobile app with the click of a button.
Whether the feature was a direct response to Skype, we may never know, but the story behind how Hangouts came to launch on Google+ is an interesting one.
For example, did you know that Hangouts took a little over an hour to be prototyped and was rolled out internally within the same day?
This information came courtesy of Google Engineering Director Chee Chew, who took the time to explain the development of the Google Hangouts feature in a recent (you guessed it) Hangout on Dan McDermott’s Google Plus Week online show.
From idea to prototype
During the Hangout, McDermott asked Chew whether the idea for Google Hangouts came from Google co-founder Sergey Brin. Whilst Brin was certainly involved in the process — by putting “some gas behind it” — Google Hangouts originally started with the idea that Google Talk should be rolled out onto the social network.
Chew was running Google Talk at the time and in a meeting with Brin and Google’s SVP of Engineering Vic Gundotra, they asked him to put chat within Google+. Chew had another idea:
“We already had this thing that was a permanent video network between Kirkland (a suburb of Seattle) and Stockholm. We were partners, but we had a hard time communicating with each other so we set up a permanent link and that link was called the Hangout. I tried to describe this to Sergey and Vic, and they kind of got it. So I opened my laptop and joined the Hangout and there were people from Kirkland and Stockholm and then all of a sudden there was Sergey and Vic. So we hung out and chatted really quickly and they asked me how long it would take for me to prototype it. So I said, it would take about and hour. We already had the code and it was just a matter of gluing it together, so Sergey said ‘Go ahead and do it, do it right now.'”
Incorporating Hangouts into Google+
Chew was meant to be taking a flight from Google HQ in Mountain View to Seattle three hours after his meeting with Brin and Gundotra, so before he left, he was “dragged over to a developer” and was instructed to tell him how to implement his Hangout code into the platform.
“We sat down for fifteen minutes,” Chew said, “He wrote the code and and I told him how to construct the link and then I took off to the airport. I called in from Seattle Airport, just to check on how things were going, to see if things were okay and whether they needed help. Whilst I was in-flight, that developer had finished the code, pushed the servers and deployed it internally. Vic was already in a hangout and he pinged me as soon as I logged in. So, I joined it and there was a Hangout with like 15 people in there. Vic shouted ‘Weeee! We have Hangouts!’ and that’s how it started.”
The original Hangout service was similar to how it looks now, it had thumbnails and a main screen but it was not nearly as polished but it was very unstable. At the time it was a work in progress.
It turns out that the Hangouts team was the very last team to join Google+, so initially people within the team didn’t understand the purpose of the tool and how it operated. Chew and his partner worried over early feedback, which labelled it as a basic video-conferencing service, but as the it rolled out publicly, the unease was dispelled as Google+ users took to it and started hosting their own Hangouts with the simple click of a button.
You can see Dan McDermott’s Hangout, complete with Chew’s history of the service, embedded below:
The Hangout, today
Google has continued to evolve the Hangouts service, adding new features and holding live events with many high-profile figures, including President Obama, the Dalai Lama and Desmond Tutu.
The service was the first piece of Google’s social layer to have a proper API, which until recently was been in “preview” mode. The Google+ team announced that not only will the Hangouts API come out of preview, it has added a new Apps section in Hangouts to support the creation of developers:
One of the most important ways we connect with others is in person. That’s why we’re so excited about Google+ Hangouts, and why we launched a preview of the Hangouts API a few months ago. Today we’re moving this API out of preview, and enabling developers to launch and share their hangout apps with the entire Google+ community! Hangout apps are regular web apps, running in a big window inside the Hangout UI. In addition to using shared-state APIs to give users real-time interactivity, you also have access to built-in Hangout features.
The company will be highlighting some of the best apps available for Hangouts, which include a poker game called “Aces Hangout”:
The Apps section will also feature all of the effects that Google+ has been building internally, such as adding an overlay onto your video, turning you into a cat or dog. It’s clear that Google feels like it has hit a sweet spot with Hangouts, as most of the newest releases within Google+ have had to do with the feature.
For example, Hangouts recently got Google Voice and Docs integration, which appeal to business users. These new apps will appeal to regular consumers who want to chat with friends or family over video. By allowing you to do more while having a video chat, Hangouts certainly appear to offer more options than Skype or FaceTime at this point.
Read next: Zeebox's Anthony Rose on the incredible rise of this social TV app, its Android app, US launch and more |
[Ct Post] - After a two-year stint in the Army, Montelli coached basketball and baseball at Sacred Heart grammar school, thanks to Monsignor John McGough, who had promised Montelli a job after his enlistment was over. The next year, he started at Notre Dame.
The group – Paul Michael Laing and John McGough, later joined by Graeme Ian Adams – eventually led the 19-year-old to the station kiosk and the police were contacted. But it resulted in the trio, plus “informant” Paul Clark, ... |
Identification of a novel family of proteins in snake venoms. Purification and structural characterization of nawaprin from Naja nigricollis snake venom.
The three-dimensional structure of nawaprin has been determined by nuclear magnetic resonance spectroscopy. This 51-amino acid residue peptide was isolated from the venom of the spitting cobra, Naja nigricollis, and is the first member of a new family of snake venom proteins referred to as waprins. Nawaprin is relatively flat and disc-like in shape, characterized by a spiral backbone configuration that forms outer and inner circular segments. The two circular segments are held together by four disulfide bonds, three of which are clustered at the base of the molecule. The inner segment contains a short antiparallel beta-sheet, whereas the outer segment is devoid of secondary structures except for a small turn or 310 helix. The structure of nawaprin is very similar to elafin, a human leukocyte elastase-specific inhibitor. Although substantial parts of the nawaprin molecule are well defined, the tips of the outer and inner circular segments, which are hypothesized to be critical for binding interactions, are apparently disordered, similar to that found in elafin. The amino acid residues in these important regions in nawaprin are different from those in elafin, suggesting that nawaprin is not an elastase-specific inhibitor and therefore has a different function in the snake venom. |
there. itwasrazortightinthe past. democrats won by .3 of the overall vote. the president is now up to 65 versus 263. what puts the president in 269? new hampshire. governor romney would have to win the only state left in our scenario on the map, and that is electoral votes in the state of iowa. that is how you get to 69 up to 269. based on the information on polling and polling data that we have, this is possible. martha: there have been folks who have been talking about the scenario for a couple of months. it could be possible. what happens then? if it is to 69 and 269, funky stuff goes on. it was the house of representatives. whoever has the majority coming on the president. whoever has the majority in the senate within vote on the vice president. how about a romney and biden ticket. after all of this? [laughter] martha: i think somebody is going to sell the story on broadway. that's just me. thank you, bill. this story has been getting a lot of attention this week as well. for attack victims are trying to get the massacre declared a terrorist attack. >> and has been nearly three |
Random forest as one-class classifier and infrared spectroscopy for food adulteration detection.
This paper proposes the use of random forest for adulteration detection purposes, combining the random forest algorithm with the artificial generation of outliers from the authentic samples. This proposal was applied in two food adulteration studies: evening primrose oils using ATR-FTIR spectroscopy and ground nutmeg using NIR diffuse reflectance spectroscopy. The primrose oil was adulterated with soybean, corn and sunflower oils, and the model was validated using these adulterated oils and other different oils, such as rosehip and andiroba, in pure and adulterated forms. The ground nutmeg was adulterated with cumin, commercial monosodium glutamate, soil, roasted coffee husks and wood sawdust. For the primrose oil, the proposed method presented superior performance than PLS-DA and similar performance to SIMCA and for the ground nutmeg, the random forest was superior to PLS-DA and SIMCA. Also, in both applications using the random forest, no sample was excluded from the external validation set. |
Bug spray, rodent poison and herbicides will be accepted along with oil-based paint. Water-based paint and latex paint will not be accepted and should be left open until they have dried and added to regular curb-side trash. |
1. Field of the Invention
The present invention relates to an image reading device which, while conveying an elongated image information support on which a plurality of frame images are recorded, reads the frame images by a line scanner.
2. Description of the Related Art
A technique is known in which frame images recorded on an image information support, e.g., a photographic film, are optically read by a reading sensor such as a CCD, image processing such as enlargement/reduction, various types of correction, and the like are carried out on the digital image data obtained by the reading, and an image is formed on a recording material by laser light which is modulated on the basis of the digital image data which has been subjected to image processing.
In this technique of digitally reading an image frame by an area sensor such as a CCD, in order to realize highly accurate image reading, the frame image is preliminarily read (so-called prescanning), reading conditions corresponding to the density or the like of the frame image are determined (e.g., the amount of light to be irradiated onto the frame image, the charge accumulating time of the CCD, and the like), and reading of the frame image is carried out again under the determined reading conditions (so-called fine scanning).
However, in the case of a photographic film provided with a magnetic recording layer (hereinafter referred to as an APS film), in addition to the image reading, additional operations for processing the magnetic information (reading and writing the magnetic information) are carried out. Accordingly, in the series of film processings, the four steps of reading magnetic information from the magnetic recording layer, writing magnetic information onto the magnetic recording layer, prescanning by using the line sensor, and fine scanning by using the line sensor, are carried out.
Here, the reading speed (the APS film conveying speed) of image reading (fine scanning) by the line scanner differs per image frame. On the other hand, from the standpoint of processing efficiency, it is desirable that fine scanning of the respective image frames and writing of magnetic information are carried out simultaneously. Accordingly, a distance, in the conveying direction of the APS film, between the write position of the magnetic information (the position of the writing head) and the image information scanning position (the reading position of the line scanner) is preferably set to be an integer multiple of the image frame pitch of the film.
However, the distance from the final frame of an APS film to the final end of the film is a predetermined length which is based on manufacturers' standards. Accordingly, in a case such as that described above in which the magnetic information writing head and the line scanner are positioned along the APS film conveying path with an interval therebetween which is an integer multiple of the image frame pitch, there is the possibility that the final image frame of the APS film cannot reach the scanning position of the line scanner, and that reading of the image information of the final image frame may not be possible. (Namely, there is the possibility that the final image frame of the APS film may only be able to reach the position of the magnetic information writing head.) |
The committee selected Mark Hollick, principal of Assabet Valley Regional Technical High School, and Jason Webster, assistant principal of The Beebe School in Malden, as finalists.
Along with serving as principal of Assabet Valley Regional Technical High School the past seven years, Hollick was the assistant principal from 2003 to 2008 and a physical education teacher from 1996 to 2003.
Webster, an assistant principal at The Beebe School since 2011, was the assistant principal at Hopkinton Middle School for four years and also taught math at the middle school and high school in Marlborough.
Hollick and Webster will interview with Superintendent Kevin Lyons and school administrators this week before visiting the middle school where they will participate in meetings with staff, parents and students.
The finalists will also meet the community during a public forum Thursday night at Quinn Middle School. The forum will be start at 6:30 p.m.
Lyons plans to choose a new principal early next week.
Daniels announced in February he would leave Hudson at the end of the school year to become the principal of Monomoy Middle School on Cape Cod.
School officials advertised the position on School Spring, an education career website.
The screening committee reviewed 33 resumes, but did not feel there were two or more qualified candidates. A a result, it suspended the search to brainstorm.
The district and New England School Development Council (NESDC) sent mailings with information on the job to all middle schools in Massachusetts, Rhode Island and Connecticut, as well as Massachusetts superintendents about the position, which yielded the 16 candidates, including Webster and Hollick.
Jeff Malachowski can be reached at 508-490-7466 or jmalachowski@wickedlocal.com. Follow him on Twitter @JmalachowskiMW. |
Accumulation of tRNAMetf on 80-S ribosomes in vitro under the influence of a Met-tRNA deacylase from rat-liver microsomes.
A Met-tRNA deacylase has been partially purified from the 0.5 M KCl wash of rat liver microsomes. In preparative sucrose gradients, the active component sediments as a single band at about 6 S, corresponding to an estimated molecular weight of 1.7 X 10(5). The deacylase is specific for Met-tRNA, without discriminating between Met-tRNA f and Met-tRNAm. Met-tRNAf bound to the initiation factor IF-MP in the ternary complex, IF-MP-GTP-Met-tRNAf, or to initiation-factor-dependent, complexes with 40-S subunits or 80-S ribosomes, is protected against deacylation. However, in the course of the initiation-factor-dependent joining of the 40-S subunit complex to 60-S ribosomal subunits, the bound Met-tRNAf is exposed to added deacylase. Under these conditions, deacylation is inhibited by GTP. The tRNAMetf remains bound and accumulates on the 80-S ribosomes. |
Section 109(d)(1) of the Clean Air Act (CAA) requires that the Agency periodically review and revise, as appropriate, the air quality criteria and NAAQS for the six "criteria" air pollutants, including ambient lead. |
Pachter: Nintendo in a world of trouble
SHARE THIS ARTICLE
The release of Wii U came two years too late to save Nintendo from a poor showing in the coming console battleground, it has been claimed.
“Nintendo’s in a world of trouble right now,” Michael Pachter told ABC News 10.
“By the time they did launch a console that stacks up really well [against PS3 and Xbox 360] the other two guys passed them by.
“Publishers are pretty excited about supporting Xbox One and PS4. They really didn’t say anything about the Wii U. We know EA has no games in development for Wii U [although it seems this could change]. If others follow suit – if you see Activision pulls support, if you see Ubisoft, you see Take-Two pull support the Wii U is a Nintendo-only gaming device which is the way they were back with the NES in 1985.
“They’re not going to sell a lot of consoles if they don’t have games like FIFA, Battlefield, Call of Duty and Grand Theft Auto V. |
Pages
Thursday, March 31, 2011
I'm no idiot
When I was eight years old I went to get my ears pierced. I let them do the first one and then declared that one was fine. It hurt and I wasn't gonna do that again. My mom had to make me do the second one, and I am glad that she did.
When I was eleven years old I steped on a broken beer bottle and had to get lots of stitches. While they were cleaning my foot, I declared that I did not need it. My foot, that is. I told them to just cut it off because it hurt. Once again my mom convinced me that I would really desire that foot later on in life, and I am glad.
When I was twenty-eight years old (yesterday) I had gum surgery on one side of my mouth. It hurt. I have decided that I don't need to do the other side. I need my mom to convince me that I really don't want to lose half my teeth and it will be worth it to go through this again.
I figured I would share the good, the bad and the funny (there is some ugly - but you don't want to see the swelling!) of my gum surgery with you - don't worry, no pictures!
The Good:
1. I have a fear of eating anything, so I will probably lose some weight.
The Bad:
1. I have a fear of eating, so I am starving and I really should get more protein - last night I actually added an egg to my mashed potatos for protein.
2. I can't exercise (except taking walks they said) for "several days"
3. Did I mention that I am starving?
The Funny:
1. You should hear me talk
2. My jaw and tongue hurt so bad that I can't tell if where the actual trauma was hurts. i am sure that once my tongue and jaw stop hurting, I will know.
I thought that the worst was over when they put in the stitches, or when the numbness wore off, but I was wrong. When I brushed my teeth this morning I learned that is the worst part!
By the way, did I mention that I hate it when doctors say general statements like, "Don't over-do it". What does that mean? Can we be exact? Should I make dinner and do the dishes? SHould I get a wheelchair? Should I take naps? Can you give me a list of do's and don'ts??
Anyway, thats all for now. I just had to write because it hurts to talk still.
I love that you're starving but too scared to eat. It's like right after you have the flu, or when you're pregnant and you could eat the whole kitchen but you'd just throw it up. Sad you have to do it in two sessions! Lame sauce.
Ohh. My sister has been through gum surgery, and though I only watched from the other side, I know it was no fun. So sorry you hurt. I once had a toenail removed and that was more painful then any of my three labor experiences. The thought still makes me cringe. I imagine what you are going through to be something like that pain, which makes me all the more sorry you are hurting. Hope you can eat again soon.
This is Mom here to explain why you need the other side done, and to tell you you ARE doing it young lady because you do not want to loose one tooth you do not have to. TRUST ME. I love you sooo much honey and my prayers are with you. It is much harder to do something a second time once you know the score. But you are tough and strong and you can do it
Psalm 127:3-5 Sons are a heritage from the Lord, children a reward from him. Like arrows in the hands of a warrior are sons born in one’s youth. Blessed is the man whose quiver is full of them. They will not be put to shame when they contend with their enemies in the gate.
Psalm 128:1-6 Blessed is every one that feareth the Lord, that walketh in his ways……thy children like olive plants round about thy table….(KJV)
Psalm 144:12 Then our sons in their youth will be like well-nutured plants, and our daughters will be like pillars carved to adorn a palace.
Luke 2:40 “And the child grew and became strong; he was filled with wisdom, And the grace of God was upon him”.
Psalm 112:1-2 Blessed is the man who fears the Lord, who finds great delight in his commands. His children will be mighty in the land; the generation of the upright will be blessed.
Isaiah 61:9 Their descendants will be known among the nations and their offspring among the peoples. All who see them will acknowledge that they are a people the Lord has blessed. |
Regulatory/effector T-cell ratio is reduced in coronary artery disease.
The protective function of regulatory T cells (Treg) has been identified in experimental atherosclerosis, but the contribution of Treg to the pathogenesis of human coronary artery disease (CAD) remains poorly understood. We investigated Treg and regulatory T-cell/effector T-cell (Treg/Teff) ratio in peripheral blood samples from CAD patients using a new strategy for precise identification of Treg. METHODS AND RESULTS: Peripheral blood samples were collected from 73 stable CAD patients (55 middle-aged CAD patients and 18 old CAD patients) and 64 controls (47 middle-aged controls and 17 young controls). CD3(+)CD4(+)FoxP3(+)T cells were divided into 3 fractions: CD45RA(+)FoxP3(low)resting Treg(Fr1), CD45RA(-)FoxP3(high)activated Treg(Fr2), and CD45RA(-)FoxP3(low)non-Treg(Fr3). CAD patients had lower percentages of Fr1 and Fr2 and higher percentages of Fr3 and CD45RA(-)Foxp3(-)Teff(Fr4+5) within the CD3(+)CD4(+)T-cell population compared to age-matched controls. Treg/Teff ratio (Fr1+2/Fr3+4+5) in CAD patients was also markedly lower than in controls (middle-aged control, 0.17±0.09 vs. middle-aged CAD, 0.10±0.05; P<0.001). The percentage of CD4(+)CD28(null)T cells within the CD4(+)T-cell population was negatively correlated with Treg/Teff ratio, excluding CD4(+)CD28(null)T cells <0.3% (r=-0.27, P<0.05). High-sensitivity C-reactive protein was also negatively correlated with Treg/Teff ratio (r=-0.22, P<0.05). CAD patients had reduced Treg and Treg/Teff ratio compared to healthy controls. The present findings may be helpful when developing immunotherapy for the prevention of CAD. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.