text
stringlengths 1
2.31M
| meta
dict |
|---|---|
Introduction {#sec1}
============
The multicolored Asian ladybeetle, *Harmonia axyridis* Pallas (Coleoptera: Coccinellidae), was introduced from Asia into Western Europe and other parts of the world to control aphid and coccid populations ([@bibr02]; [@bibr10]). In Belgium, *H. axyridis* was used as a biological control agent beginning in 1997 and was first observed in the wild in 2001. Since then *H. axyridis* populations have increased and gradually expanded into Belgium ([@bibr02]).
This species presents all the characteristics shared by an efficient aphid predator: large body size, high voracity, high predation efficiency ([@bibr35]), high colonization aptitude ([@bibr60]), rapid development, high fecundity, and low susceptibility to pathogens or natural enemies ([@bibr41]). *Harmonia axyridis* has become ubiquitous in many parts of the world, including America, Europe, and Africa ([@bibr37]; [@bibr12]), and has been reported in many different habitats, such as agroecosystems, gardens, and arboreal habitats ([@bibr40]).
Due to its large body and efficient physical and chemical defenses, *H. axyridis* has become an intraguild predator ([@bibr50]; [@bibr58]). Intraguild predation has been observed among other ladybeetle species ([@bibr45]; [@bibr58]); other aphid natural enemies, including syrphids, chrysopids, and parasitoids ([@bibr46]; [@bibr59]; [@bibr31]); and aphid pathogenic fungus ([@bibr45]). This intraguild predation behavior is thought to have led to a decrease in native species ([@bibr09]; [@bibr23]; [@bibr57]; [@bibr11]; [@bibr48]). In Belgian urban areas, Adriaens et al. ([@bibr03]) found a significant decline in native species, including the two-spot ladybird, *Adalia bipunctata* L. (Coleoptera: Coccinellidae), and the 10 spotted ladybird, *Adalia decempunctata* L., on pine, lime, and maple trees following the arrival of *H. axyridis.* The decline of native species can possibly be explained by the decline in number of their principal prey, resulting in reduced survivorship in local habitats and altered dynamics of habitat use and dispersal ([@bibr21]).
According to previous reports, the most dominant aphidophage in Belgian agroecosystems appear to be two coccinellids, the seven-spotted ladybug, *Coccinella septempunctata* L. (Coleoptera: Coccinellidae), and *H. axyridis*; one hoverfly, the marmalade hoverfly, *Episyrphus balteatus* De Geer (Diptera: Syrphidae); and one braconid, the parasitic wasp *Aphidius ervi* Haliday (Hymenoptera: Braconidae) ([@bibr20]; [@bibr02]; [@bibr05]). In arboreal habitats, four coccinellids were reported as abundant species: *A. bipunctata*.,the 14-spotted ladybird, *Propylea quatuordecimpunctata* L. (Coleoptera: Coccinellidae), the 22-spot ladybird *Psyllobora vigintiduopunctata* (L.), and *H. axyridis* ([@bibr02]). In 2001, the same results were observed by Francis ([@bibr22]), with the exception of *H. axyridis.*
Our study was conducted eight years after the first observation of *H. axyridis* in the wild in Belgium ([@bibr01]). Following aphidophagous decline highlighted by several studies, the current study was conducted in order to assess the relative abundance of *H. axyridis* through the quantification of the abundance of this exotic species and other aphidophages in four important Belgian crops (wheat, *Triticum aestivum* L. (Poales: Poaceae), corn, *Zea mays,* potato, *Solanum tuberosum* (Solanales: Solanaceae), and broad bean *Vicia faba* (Fabales: Fabaceae)) using a three-year inventory.
Materials and Methods {#sec2}
=====================
Study site {#sec2a}
----------
Aphidophagous insect populations were sampled from 2009 to 2011 in Hesbaye, an intensive agricultural production area in Wallonia, the southern region of Belgium (individual sites given in [Table 1](#t01_01){ref-type="table"}). Four crops were chosen for their agronomic importance: wheat, corn, potato, and broad bean *Vicia faba.* The sampling period ran from mid-May to late September. Every week, nine fields of each crop were sampled.
Sampling methods {#sec2b}
----------------
The sampling methods used to assess the numbers of aphidophagous predators and aphids consisted of whole-plant visual inspections, using 1 m^2^ quadrats distributed randomly throughout the whole field. In order to avoid the influence of surrounding crops, a 20 m buffer zone around the edge of each field was not sampled. Visual sampling was conducted, as it provides an easy and accurate method for the estimation of larval and adult densities of coccinellids in agroecosystems ([@bibr43]). Forty-eight quadrats were examined in each crop every week. Quadrats were located along transect lines across each field and spaced 20 m apart. All leaves and stems within the quadrat were examined, and all aphidophagous stages were recorded. Aphid species were also quantified on all leaves and stems. Larvae and pupae were brought to the laboratory to develop under laboratory conditions (24 ± 1° C, 75 ± 5% RH) for identification at the species level. All aphid predators were identified, with the exception of members of the common green lacewing, *Chrysoperla carnea* Stephens (Neuroptera: Chrysopidae) species complex, which were pooled together. This group comprises three cryptic species, *C. kolthoffi* Navas, *C. lucasina* Lacroix, and *C. carnea* Stephens, which can only be differentiated using molecular techniques ([@bibr07]; [@bibr38]).
Statistical analysis {#sec2c}
--------------------
In order to assess the species diversity, the collected data were analyzed by considering (1) species richness and (2) evenness according to the Shannon diversity index (H) ([@bibr39]), which considers both the number of species and the distribution of individuals among species. H is minimal if all individuals belong to only one species or if all species are represented by one individual; H is maximal if all individuals are evenly distributed. Evenness (J) varies from 0, if only one species dominates, to 1, if all species show similar abundance. Evenness (J) and the Shannon diversity index (H) were calculated as follows:
Because mean densities per m^2^ were low, these values are presented per 100 m^2^ The mean abundance per species was analyzed with an analysis of variance (ANOVA: general linear models) with crops (q = 4) and years (n = 3) used as factors (α = 0.05). Within crops, densities of species were compared using the least square difference (LSD; α = 0.05). To account for the variations of predator abundances, the abundance per species was analyzed using an analysis of covariance (ANCOVA: general linear models) with crops (q = 4) and years (n = 3) used as factors (α = 0.05) and aphid densities used as the covariable. Prior to analyses, a log~10~(x + 1) was used to transform the data distribution (counting) due to its asymmetry ([@bibr19]). Although statistical analyses were performed on transformed data, untransformed data are presented in [Tables 4](#t04_01){ref-type="table"} and [5](#t05_01){ref-type="table"}. Statistical analyses were performed using Minitab® release 1.5 ([www.minitab.com](www.minitab.com)).
Results {#sec3}
=======
Diversity of aphidophages {#sec3a}
-------------------------
During the three years (2009--2011), 11 aphidophagous taxa were observed on the four different cultures: the hoverfly *E. balteatus,* the coccinellids *C. septempunctata, C. quinquepunctata* L. (Coleoptera: Coccinellidae), *C. undecimpunctata* (L.), *H. axyridis, P. quatuordecimpunctata, A. bipunctata, A. decempunctata,* the cream-spot ladybird, *Calvia quatuordecimguttata* (L.), *Hippodamia variegata* Goeze; and the *C. carnea* species group ([Table 2](#t02_01){ref-type="table"}).
From 2009 to 2011, species richness increased in broad bean, corn, and wheat, reaching 6, 8, and 7 species respectively ([Table 3](#t03_01){ref-type="table"}). Species richness did not evolve in potato crops, remaining at five species during the entire period. Overall, five aphidophages were continually observed during the three-year period in each crop and represented 95% of all the aphidophage observed in 2009 and 99% in both 2010 and 2011: *E. balteatus, C. carnea* s.l., *C. septempunctata, P. quatuordecimpunctata,* and *H. axyridis.*
The evenness value was low in wheat with J = 0.35, 0.25, and 0.38 in 2009, 2010, and 2011, respectively, with the numerical dominance of two species: *C. septempunctata* and *E. balteatus* ([Tables 3](#t03_01){ref-type="table"}, [4](#t04_01){ref-type="table"}, and [5](#t05_01){ref-type="table"}). In broad bean and potato, evenness varied from year to year. In broad bean, *C. septempunctata* was the most abundant in 2010 (J = 0.45) and 2011 (J = 0.59). In potato, *C. carnea* s.l. and *C. septempunctata* numerically dominated the aphidophagous guild in all three years (J = 0.65, 0.79, 0.74) ([Table 3](#t03_01){ref-type="table"}). In corn, the evenness during the three years decreased from 0.82 (2009) to 0.61 (2010).
Relative abundance of aphidophage in four crops {#sec3b}
-----------------------------------------------
The relative abundance of both adult and larval stages of the five aphidophage within each crop showed significant differences ([Tables 4](#t04_01){ref-type="table"} and [5](#t05_01){ref-type="table"}), with the exception of potato in 2010, in which densities of larvae populations of different predators were not significantly different *(F~4.3570~* = 1.4*; p* = 0.25) ([Table 5](#t05_01){ref-type="table"}).
Corn {#sec3c}
----
The most abundant adult species in 2009 and 2010 was *C. carnea* s.l. and *C. septempunctata* (2009: *F~4,1065~* = 10.7, *p* \< 0.001 ; 2010: *F~4,3425~* = 14.3, *p* \< 0.001) *(p* \< 0.05; LSD) ([Table 4](#t04_01){ref-type="table"}). In 2011, *H. axyridis* and *C. septempunctata* densities were both significantly higher than those of other species *(p* \< 0.05; LSD) *(F~4,4055~* = 24.2, *p* \< 0.001) ([Table 4](#t04_01){ref-type="table"}). *H. axyridis* larvae densities were significantly higher in 2009 and 2011 (2009: *F~4,1065~* = 5.4, *p* \< 0.001 ; 2010: *F~4,3425~* = 25.5, *p* \< 0.001 ; 2011: *F~4,4055~* = 57.0, *p* \< 0.001) than those observed for the other species ([Table 5](#t05_01){ref-type="table"}).
Wheat {#sec3d}
-----
Adult populations of *H. axyridis* remained lower than other species during the three-year sampling (2009: *F*~*4,1765*~ = 5.4, *p* \< 0.001 ; 2010: *F~4,2925~* = 24.1*, p* \< 0.001; 2011: *F~4,2625~* = 38.0, *p* \< 0.001) *(p* \< 0.05; LSD) and did not exceed 1.1 ± 0.6 adults per 100 m^2^ ([Table 4](#t04_01){ref-type="table"}). *Episyrphus balteatus* was the most frequently encountered adult species in 2009, whereas in 2010 and 2011 *C. septempunctata* was the most abundant species *(p* \< 0.05; LSD) ([Table 4](#t04_01){ref-type="table"}). Larvae of *E. balteatus* were the most abundant *(p* \< 0.05; LSD) of the aphidophage during the three years (2009: *F~4,1765~* =11.7, *p* \< 0.001 ; 2010: *F~4,2925~ =* 91.1*, p* \< 0.001; 2011: *F*~*4,2625*~ = 213.6, *p* \< 0.001) ([Table 5](#t05_01){ref-type="table"}).
Potato {#sec3e}
------
Trends for *H. axyridis* were the same in potato as in wheat: *H. axyridis* was not the most abundant species, and its density did not exceed 2.5 ± 0.7 adults per 100 m^2^ ([Table 4](#t04_01){ref-type="table"}). Two species were more abundant than others: *C. carnea* s.l. in 2009 and 2011 (2009: *F*~*4,500*~ = 7.7, *p* \< 0.001 ; 2011: *F~4,3250~* = 10.8, *p* \< 0.001), and *C. septempunctata* in 2010 *(F~4,3570~* = 12.8, *p* \< 0.001) and 2011 *(p* \> 0.05; LSD). In 2009, larvae of *C. carnea* s.l., *P. quatuordecimpunctata,* and *H. axyridis* were the most abundant species but densities remained low ([Table 5](#t05_01){ref-type="table"}). In 2010, larvae densities of the five above-mentioned species were not significantly different from each other *(F~4,3570~* = 1.4, *p* = 0.25) ([Table 5](#t05_01){ref-type="table"}). In 2011, only *C. carnea* s.l. numerically dominated the aphidophages community (*F~4,3250~* = 5.5, *p* \< 0.001) *(p* \> 0.05; LSD) ([Table 5](#t05_01){ref-type="table"}).
Broad bean {#sec3f}
----------
In 2009, *H. axyridis* and *C. septempunctata* adults were the only adult species observed in broad bean (*F*~*4,160*~ = 0.7, *p* \> 0.05). *Coccinella septempunctata* was the most abundant species in 2010 *(F~4,2415~* = 22.1, *p* \< 0.001), while in 2011 both *C. septempunctata* and *C. carnea* s.l. were profusely observed (*F~4,280~* = 37.5, *p* \< 0.001) *(p* \< 0.05; LSD) ([Table 4](#t04_01){ref-type="table"}). In 2009, three species were present at the larval stage: *P. quatuordecimpunctata, C. carnea* s.l., and *E. balteatus* (*F~4,160~* = 1.0 *,p* = 0.43) *(p* \< 0.05; LSD). In 2010 (*F~4,2415~* = 8.35*, p* \< 0.001) and 2011 *(F~4,2480~* = 21.2, *p* \< 0.001), all species were observed, and *C. septempunctata* was the most abundant ([Table 5](#t05_01){ref-type="table"}).
Effect of aphid densities and sampling year on relative abundance of aphidophage {#sec3g}
--------------------------------------------------------------------------------
Abundances of *H. axyridis* in wheat and potatoes were not analyzed, due to very low numbers of individuals observed during the three-year inventory. Over the three-year sampling, adult populations of *H. axyridis* in corn significantly increased (*F*~2,1709~ = 14.5, *p* \< 0.001) ([Table 6](#t06_01){ref-type="table"}) from 7.8 ± 2.0 in 2009 to 19.9 ± 1.8 individuals per 100 m^2^ in 2011 ([Table 4](#t04_01){ref-type="table"}). Larval populations in the same crop also increased statistically (*F*~2,1709~ = 39.9, *p* \< 0.001) ([Table 6](#t06_01){ref-type="table"}), rising from 11.9 ± 4.2 to 70.8 ± 6.6 larvae per 100 m^2^ ([Table 5](#t05_01){ref-type="table"}). In broad bean, relative abundance of *H. axyridis* was not significantly different among the three years, neither at the adult ([Table 6](#t06_01){ref-type="table"}) nor larval ([Table 7](#t07_01){ref-type="table"}) stages.
*Coccinella septempunctata* larvae declined in broad bean (F~2,1011~ = 4.7, *p* = 0.009), wheat (*F*~2,1463~ = 14.4, *p* \< 0.001), and corn (*F*~2,1709~ = 31.9, *p* \< 0.001) ([Table 7](#t07_01){ref-type="table"}); densities decreased by 10 and 29.6 times in corn and broad bean respectively. In wehat, no larvae were observed in 2011, while 13.4 ± 3.9 larvae per 100 m^2^ were observed in 2009.
The abundances of three other aphidophage showed variable changes ([Tables 6](#t06_01){ref-type="table"} and [7](#t07_01){ref-type="table"}).
Aphid populations and correlation with aphidophage densities {#sec3h}
------------------------------------------------------------
In 2009, 2010, and 2011, seven, nine, and 10 species of aphids were identified, respectively: the pea aphid, *Acyrthosiphon pisum* Scopoli (Hemiptera: Aphididae); the cowpea aphid, *Aphis craccivora* Koch; the black bean aphid, *Aphis fabae* Scopoli; *Aphis frangulae* Walker; the buckthorn aphid, *Aphis nasturtii* Kaltenbach; the potato aphid, *Macrosiphon* *euphorbiae* Thomas; the vetch aphid, *Megoura viciae* Buckton; the rose grain aphid, *Metopolophium dirhodum* Walker; the green peach aphid, *Mizns persicae* Sulzer; *Rhopalosiphum* sp.; and *Sitobion* sp. ([Table 8](#t08_01){ref-type="table"}). The mean number of observed aphids increased in corn (*F~2,2589~* = 39., *p* \< 0.001) and potato (*F~2,1410~* = 17.11, *p* \< 0.001) from 2009 to 2011. Aphid densities also statistically varied in broad bean (*F~2,974~* = 8.7, *p* \< 0.001) and wheat (*F~2,1392~* = 102.7, *p* \< 0.001) from 2009 to 2011, but without any general evolution ([Table 8](#t08_01){ref-type="table"}).
The ANCOVA analyses showed a linear relationship between aphid and predator populations in 55% of adult populations ([Table 6](#t06_01){ref-type="table"}) and 35% of larvae populations ([Table 7](#t07_01){ref-type="table"}) (*p*~aphids~ \< 0.05, ANCOVA). In these cases, aphid densities influenced the predator abundance. Results (*p*~years~) comparison between ANOVA and ANCOVA showed that the influence of aphid populations on predator abundance variations between years was not statistically significant.; *p~years~* of the two statistic analyses showed the same results.
Discussion {#sec4}
==========
Since the invasive coccinellid *H. axyridis* spread over Europe ([@bibr10]), imposing negative impacts on native aphidophage and affecting composition of several guilds ([@bibr53]; [@bibr48]), studies have evaluated the population spread of this coccinellid. The current study\'s sampling of aphid predators in Belgian agroecosystems from 2009 to 2011 showed that *H. axyridis* lives and reproduces more efficiently in corn and broad bean than in wheat and potato. In corn, the evenness during the three years decreased when *H. axyridis* population increased strongly and was higher than the population of other species.
During the three-year sampling, 11 aphidophage were observed in these agroecosystems, but five of them predominated: *E. balteatus, C. septempunctata, P. quatuordecimpunctata, H. axyridis,* and *C. carnea* s.l. Five dominant species in agroecosystems is a common observation ([@bibr26]). Observations on predator densities highlight that *H. axyridis* was not the numerically dominant species in every crop: in wheat and potato, *C. septempunctata* was more abundant than *H. axyridis.* In many European agricultural crops, both *C. septempunctata* and *P. quatuordecimpunctata* were dominant prior to the arrival of *H. axyridis* (Honěk 1979; [@bibr06]; [@bibr14]), and it appears that these two species have maintained their dominance in spite of being prone to intraguild predation by *H. axyridis* in the field ([@bibr24]).
In our study, larvae of *E. balteatus* were the most abundant observed predators in wheat, which has already been reported by Tenhumberg and Poehling ([@bibr55]) prior to the arrival of *H. axyridis. Episyrphus balteatus* has also been previously reported as one of the most abundant aphidophage in vegetable crops, such as broad beans ([@bibr17]; [@bibr16]). This could be explained by abiotic conditions (high density cereal crop, with high humidity and low temperature) that are more favorable to the larvae of *E. balteatus* (Honěk 1983).
The fact that *H. axyridis* is not the most abundant aphidophage in agrosystems is probably due to its generalist behavior and arboreal habitat selection ([@bibr25]; [@bibr15]; [@bibr36]; [@bibr09]; [@bibr34]). However, it has been reported that *H. axyridis* can also thrive in agrosystems such as wheat, corn, and potato ([@bibr36]; [@bibr13]; [@bibr16]; [@bibr42]; [@bibr08]; [@bibr44]; [@bibr50]; [@bibr32]), as well as in herbaceous habitats ([@bibr36]; [@bibr33]; [@bibr04]).
There were evident changes in the abundance of aphidophages in crops through the years, but this study does not propose to identify a global evolution (increase or decline) in any of the species that were observed in this study. The causes of such fluctuations are diverse and may include factors such as the diversity and abundance of aphid species ([@bibr61]; Honěk 1982; [@bibr56]). The results of the ANCOVA showed that there was a linear relationship between prey and predator populations, but additional biotic and abiotic factors contribute to the annual variability of predator abundance. Climate could be one such factor, due to its influence on natural enemies, overwintering mortality, and aphid populations ([@bibr26]; Szentkirályi 2001; [@bibr47]). Several other factors could also explain the variation between crops: insolation, humidity (Honěk 1985), quantity and quality of host plants ([@bibr05]), and adjacent habitats ([@bibr17]; [@bibr05]).
A particularly interesting finding is that although *H. axyridis* breeding occurred in all four inventoried crops to some extent, adults of this species are not ubiquitous; few immature individuals were recovered from potato and wheat. Assuming that declines in native species are caused by *H. axyridis* ([@bibr48]), this suggests that certain crops, such as wheat and potato, could act as refuges from *H. axyridis* at certain times, while native species, such as *E. balteatus* and *C. septempunctata,* are able to breed with a lower risk of intraguild predation or other forms of competition from the invaders. Such habitats could become even more important as native species adapt to the invader by evolving to avoid habitats where *H. axyridis* occurs in high numbers, as has been seen in co-occurring aphidophages in their native habitats ([@bibr51]).
In conclusion, our study indicates that *H. axyridis* was not the most frequently observed aphidophage in the four most important Belgian agronomical crops. In future studies, longer samplings would be preferable in order to eventually identify quantitative changes in the native fauna suggested from other studies. Agroecosystems may even constitute an ecological reservoir for certain native aphidophage.
######
Sites surveyed for aphidophage from 2009 to 2011 in Belgium.
######
Aphidophage diversity in four crops (broad bean, wheat, corn, and potato) from 2009 to 2011.
######
Species richness and diversity index (H= Shannon-Weiner diversity index, where absolute diversity = 1.00; J = evenness or relative diversity (H/Hmax), where absolute evenness = 1.00).
######
Abundance (means and SE) of aphidophage (adults) per 100 m^2^ in four agroeco system s from 2009 to 2011. Means within a crop followed by the same letter were not significantly different for the same field (*p* \> 0.05; LSD test).
######
Abundance (means and SE) of aphidophage (larvae) per 100 m^2^ in four agroecosystems from 2009 to 2011. Means within a crop followed by the same letter were not significantly different for the same field (*p* \> 0.05; LSD test).
######
ANOVA and ANCOVA summary of the effects of aphid abundance and year sampling (2009, 2010, 2011) on the abundane of five predators at the adult stage in four crops (corn, wheat, potato, and broad bean).
######
ANOVA and ANCOVA summary of the effects of aphid abundance and year sampling (2009, 2010, 2011) on the abundance of five predators at the larval stage in four crops (corn, wheat, potato, and broad bean).
######
Mean numbers and SE of aphids/100 m^2^ observed in four fields (wheat, broad bean, corn, and potato) from 2009 to 2011.
We thank V. Sibret and A.M. Buset for their technical assistance, Dr Y. Brostaux for his advice on statistical analysis, and P. Leroy and C. De Clerck for their helpful comments on previous versions of the manuscript. The authors are grateful to Professor B. Bodson and F. Vancutsem of the Unité de Phytotechnie des Régions Tempérées (ULg, GxABT) for the availability of fields. This research was funded by the Service Public de Wallonie (SPW - DGO3, project n°D311197). Delphine Durieux was financially supported by a Ph.D. grant from the Fonds pour la formation à la Recherche dans l\'Industrie et l\'Agriculture (FRIA), Belgium.
[^1]: **Editor:** Xinzhi Ni was editor of this paper.
|
{
"pile_set_name": "PubMed Central"
}
|
/* global hexo */
'use strict';
var assign = require('object-assign');
hexo.config.index_generator = assign({
per_page: typeof hexo.config.per_page === 'undefined' ? 10 : hexo.config.per_page,
order_by: '-date'
}, hexo.config.index_generator);
hexo.extend.generator.register('index', require('./lib/generator'));
|
{
"pile_set_name": "Github"
}
|
Incidence of fracture through metastases in long bones.
A knowledge of the incidence of fractures through metastases of various sizes in long bones is important when considering prophylactic internal fixation. The notes and X-rays of 66 consecutive patients with 100 metastases in long bones were analysed. Fractures are highly unlikely to occur (2.3 per cent) when less than 50 per cent of the cortex is destroyed and most likely to occur (80 per cent) when over 75 per cent of the cortex is destroyed.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Q:
Read selected text in vim
I have the following bash script (play.bash):
#!/bin/bash
pico2wave -l=en-US -w=/tmp/test.wav "$1"
aplay /tmp/test.wav
rm /tmp/test.wav
I would like to use it for reading selected text in the vim editor.
The following does not work
:'<,'>w !bash play.bash
The command executes, but I'm not hearing it.
I'd like to only hear the text, without leaving the vim window.
Is this possible?
A:
When using the ! command in vim, your selection will be piped to your script on standard input, not given as a command line argument.
A simple solution would be to modify your script so it reads from standard input:
xargs pico2wave -l=en-US -w=/tmp/test.wav
xargs reads from standard input, then calls pico2wave with the data as command line arguments, removing newlines.
|
{
"pile_set_name": "StackExchange"
}
|
of -4*f + 22*f + 3*f + 9*f - 7*f**2 wrt f.
-14
Let v(u) = 3*u + 1. Let h(n) = -3*n - 3. Let a(t) = -3*t - 2. Let c(l) = -3*a(l) + 2*h(l). Let s(g) = 3*c(g) - 2*v(g). Differentiate s(z) with respect to z.
3
Let g(m) be the first derivative of m**2 - 1/12*m**4 + 0*m + 1/3*m**3 - 2. Let c(s) be the second derivative of g(s). Differentiate c(q) with respect to q.
-2
Let o(n) = -4*n**3 + 2*n**2 - 1. Let b be o(1). Let a be 0 + 2/(-3)*b. Differentiate -a + 4 + 0 - h with respect to h.
-1
Let z = 3 - 1. Suppose 5*a - 2*l - l - 11 = 0, 3*l = -4*a - 2. What is the first derivative of -z*d**3 + 3*d**3 - a + 0 wrt d?
3*d**2
Let b be (-9 - -1)/(2/4). Let f be (b/12)/(4/(-6)). Find the third derivative of 0*d**2 - 2*d**f - d**2 + 4*d**4 wrt d.
96*d
Let d(n) be the third derivative of 3*n**5/20 - n**4/3 + 7*n**2. What is the second derivative of d(i) wrt i?
18
Find the second derivative of 0*h**3 + 7*h - 4*h**3 + 8*h**3 wrt h.
24*h
Suppose -4*a + 5*o + 31 = -a, 5*a + o = 5. Find the second derivative of m**a - 2*m**3 - m**2 - m**3 + m wrt m.
-18*m
Let p(h) = -8*h**2 - 14. Let c(o) = -4*o**2 - 7. Let z(n) = 5*c(n) - 3*p(n). What is the first derivative of z(g) wrt g?
8*g
Suppose 22 - 7 = 3*n. Let r be ((-1)/(-3))/(n/15). What is the first derivative of r - 1 + 3 - t wrt t?
-1
Suppose 4*i = x - 19, 5*x = 5*i + 15 + 20. Find the second derivative of x*f + 2 - 3*f**4 - 6*f - 2 wrt f.
-36*f**2
Let l(t) = -t + 4. Let y be l(4). Suppose -2*k = -k - d, 2*d = 5*k - 6. Find the third derivative of v**2 + v**k + y*v**5 - 3*v**2 + 3*v**5 wrt v.
180*v**2
Let g be 2*5*(-4)/(-8). Differentiate -5*x + 0*x - 8 - g*x + 3*x wrt x.
-7
Suppose -10 = 4*p + 10. Let g = p - -9. What is the second derivative of u - 4*u - 3*u**4 + u**g wrt u?
-24*u**2
Let n be (-6)/(-4) - 2/(-4). Suppose n*p - 3 = -p. Find the first derivative of -p - 2*c + 4 + 0 wrt c.
-2
What is the second derivative of -9*l**4 - 5*l - 3*l + 10*l wrt l?
-108*l**2
Let o be (1/(-2))/1*54. Let y be (-8)/20 - o/5. Find the second derivative of 2*h**4 + 3*h - 2*h**4 + 3*h**y + h**5 wrt h.
80*h**3
What is the third derivative of 20 - 17*i**2 - 40 + 20 + 6*i**6 wrt i?
720*i**3
Let k(n) be the first derivative of 3*n**4/4 + 3*n**2/2 + 9. What is the second derivative of k(m) wrt m?
18*m
Suppose 6 = 2*y - 28. Let s = y - 12. What is the third derivative of s*t**6 + 4*t**6 - 8*t**6 - t**2 wrt t?
120*t**3
What is the second derivative of -13*i**2 + 40*i + 46*i**2 - 28*i wrt i?
66
Let k(m) = 19*m**3 + 6*m**2 + 18*m + 6. Let i(a) = 38*a**3 + 13*a**2 + 36*a + 13. Let n(b) = 6*i(b) - 13*k(b). What is the second derivative of n(g) wrt g?
-114*g
What is the derivative of 5 + 4 + k**2 - 1 + 0*k**2 wrt k?
2*k
Let h(k) be the third derivative of 0 - 5/336*k**8 + 0*k + 6*k**2 + 0*k**7 + 1/15*k**5 + 0*k**3 + 0*k**6 + 0*k**4. What is the third derivative of h(c) wrt c?
-300*c**2
Suppose -17*z + 13 - 13 = 0. Let l(p) be the first derivative of 1/2*p**2 + p**4 + 0*p - 1 + z*p**3. Find the second derivative of l(u) wrt u.
24*u
Let u(o) be the first derivative of -45*o**4/4 - o**3/3 - o**2/2 - 4*o - 59. Find the second derivative of u(y) wrt y.
-270*y - 2
Let j(b) be the first derivative of -b**7/35 - b**5/20 - 5*b**2/2 - 6. Let d(s) be the second derivative of j(s). Find the third derivative of d(i) wrt i.
-144*i
Let s be (2/(-4))/((-1)/12). Find the third derivative of -4*q**6 + q**6 + q**s + q**2 wrt q.
-240*q**3
Let b = -3 - -5. Let p(v) be the second derivative of 1/12*v**4 + 2*v + 0*v**b + 1/6*v**3 + 0. What is the second derivative of p(k) wrt k?
2
Let u(h) be the second derivative of 0*h**6 + 0*h**5 + 3/28*h**8 + 0*h**2 + 0*h**3 + 0 + 7*h + 0*h**7 - 5/12*h**4. Find the third derivative of u(t) wrt t.
720*t**3
What is the third derivative of -4*y**3 - 21*y**6 + y**2 + 58*y**6 + 4*y**3 wrt y?
4440*y**3
Let f(o) = 8*o + 6. Let g(c) = -c - 1. Let n(v) = v**3 - 4*v**2 + 2*v - 3. Let w be n(4). Let p(l) = w*g(l) + f(l). Differentiate p(q) wrt q.
3
Differentiate 9*c**2 + 31 - 9*c**2 + 0*c**2 - 7*c**3 wrt c.
-21*c**2
Let n(h) = -10*h**4 - 3*h**3 - 3*h**2 + 4. Let a(z) = z**4 + z**3 + z**2 - 1. Let r(b) = -3*a(b) - n(b). What is the first derivative of r(g) wrt g?
28*g**3
Suppose 2*z + 2 = 34. Suppose -4 + z = 4*q. What is the second derivative of -u + 2*u**q - 2*u**3 - u**3 wrt u?
-6*u
Let b(t) be the third derivative of 0*t**7 + 0*t**4 + 0*t + 0*t**6 + 0*t**3 + 0 - 3*t**2 + 1/60*t**5 - 1/168*t**8. Find the third derivative of b(d) wrt d.
-120*d**2
Suppose -7*t = -11*t + 12. Differentiate -5 + t*u + 5*u - 2*u - 2*u with respect to u.
4
Let v(n) = 10*n**4 - 5*n**2 + 5*n - 16. Let q(u) = 11*u**4 - 6*u**2 + 6*u - 17. Let y(h) = -5*q(h) + 6*v(h). What is the first derivative of y(w) wrt w?
20*w**3
Let p(k) be the first derivative of 2*k**5/5 - 4*k - 3. Differentiate p(y) with respect to y.
8*y**3
Let x be (-9)/(-12)*(-4)/(-1). Find the second derivative of 2*i**3 - i - 2*i**3 + 2*i**3 - x*i wrt i.
12*i
Find the third derivative of -31*a**4 + 13*a**4 - 28*a**4 - 2*a - 9*a**2 - a**4 wrt a.
-1128*a
Let i be (-1 + 1)/(3 - 1). Let x(c) be the second derivative of -1/10*c**5 + 0*c**3 + 0 - c**2 - 2*c + i*c**4. What is the first derivative of x(l) wrt l?
-6*l**2
Let z(h) be the second derivative of 4*h**6/5 + 5*h**3/2 - 19*h. What is the second derivative of z(d) wrt d?
288*d**2
Let t(s) be the second derivative of 1/20*s**5 + 2*s + 0 + 0*s**3 + 1/2*s**2 + 0*s**4. Differentiate t(w) with respect to w.
3*w**2
Let p be (-1)/(-1) - (6 + -8). Suppose -t + 4*n = -13 - 7, 0 = -p*t + 4*n + 28. Find the third derivative of -k**t - 2*k**5 - 3*k**2 + k**4 + 5*k**5 wrt k.
180*k**2
Let d(l) = l**2 - 7*l + 8. Let h be d(6). What is the third derivative of 54*o**3 + 2*o**2 - 50*o**3 + h*o**2 wrt o?
24
Let h(d) = 2*d**3 + 5 - d**3 - 4*d**2 - 1. Let c be h(4). What is the derivative of -3 - 2*b**c + 4*b**4 - b**4 wrt b?
4*b**3
Let h be 8/6 - 2/(-3). Suppose -6 = -4*p + h. Find the third derivative of 0*y**6 + 5*y**2 - p*y**2 - 2*y**6 + y**6 wrt y.
-120*y**3
Let s be (1/1)/(5/10). What is the third derivative of -3*w**3 + 2*w**3 + 6*w**3 - 3*w**s wrt w?
30
Let p(c) = -c**3 + 5*c**2 + c - 3. Let b be p(5). Let f be ((-4)/6)/(6/(-45)). What is the second derivative of -t - b*t**5 + 5*t - f*t wrt t?
-40*t**3
Let r(u) be the first derivative of -16*u**5/5 - 16*u**3/3 - 9. Find the third derivative of r(l) wrt l.
-384*l
Let p(s) = s**5 - s**3 - s**2 - s + 1. Let c(l) = -11*l**5 - 3*l**3 - 3*l**2 - 32*l + 3. Let f(w) = -c(w) + 3*p(w). Find the second derivative of f(h) wrt h.
280*h**3
Let v(y) = -21*y**4 + 11*y**2 - 11*y - 66. Let p(q) = 4*q**4 - 2*q**2 + 2*q + 13. Let d(f) = 11*p(f) + 2*v(f). Differentiate d(z) with respect to z.
8*z**3
Let m(r) = -r - 12. Let n be m(-14). Find the second derivative of s + 6*s**3 + 2*s - n*s wrt s.
36*s
Suppose 2*m = -2*m + 16. Suppose 0 = -m*g + 18 - 2. What is the second derivative of 2*d + g*d**2 - 4*d**2 - 2*d**2 wrt d?
-4
Let f(d) be the third derivative of -d**6/3 + 5*d**4/8 + 12*d**2. What is the second derivative of f(b) wrt b?
-240*b
Suppose -2*f - 5*l - 16 = 0, 0*f + f - l - 6 = 0. What is the second derivative of 2*i**3 - 5*i**3 + f*i + 0*i wrt i?
-18*i
Let q = 7 + -7. Find the second derivative of q*i**2 - 4*i**2 - i - 3*i**2 + 3*i**2 wrt i.
-8
Let s(t) = 8*t**2 - 25*t. Let d(w) = -2*w**3 + 9*w**2 - 26*w. Let c(z) = -6*d(z) + 7*s(z). Find the second derivative of c(l) wrt l.
72*l + 4
Let x(y) be the second derivative of -y**6/240 + y**5/40 + y**4/3 - 3*y. Let t(n) be the third derivative of x(n). What is the first derivative of t(w) wrt w?
-3
Let m = 3 - 0. What is the derivative of -m*o + 4 + 3 - 1 - 1 wrt o?
-3
Let n(p) be the third derivative of p**8/336 - p**7/10 + p**5/60 - 5*p**4/8 - 42*p**2. What is the third derivative of n(q) wrt q?
60*q**2 - 504*q
Let r(h) be the third derivative of 31*h**4/24 - 37*h**3/6 - 5*h**2. What is the derivative of r(p) wrt p?
31
Suppose 0 = 4*k - 3 - 5, -4*k = 2*d - 12. Suppose 0 = l - 0*l - d. What is the second derivative of l*r - 2*r - 2*r**2 - 2*r wrt r?
-4
Let n(r) = -5*r**3 - 3*r**2 - 6*r. Let x(c) = c**2. Let u(d) = 2*n(d) + 6*x(d). Find the second derivative of u(q) wrt q.
-60*q
Let l(q) be the second derivative of q**6/72 - q**4/4 + q**3/6 - q. Let v(d) be the second derivativ
|
{
"pile_set_name": "DM Mathematics"
}
|
Moorthorpe railway station
Moorthorpe railway station serves Moorthorpe and South Kirkby, near Pontefract, in the City of Wakefield district of West Yorkshire, England. It lies on the Wakefield Line and the Dearne Valley Line, north of Sheffield and is served by Northern.
The station was opened in May 1879, jointly by the Midland Railway and North Eastern Railway, as part of their Swinton and Knottingley Joint Railway scheme. A short chord was also built at the same time to link the S&K Joint line with the main Doncaster to Leeds line at South Kirkby Junction. Upon opening, Moorthorpe marked the northern limit of the part of the line maintained by the Midland Railway; the southern end of the intersection bridge over the Doncaster to Leeds line, marked the actual boundary. North of that point, the North Eastern Railway undertook the maintenance. Moorthorpe station is of North Eastern Railway design, as is Pontefract Baghill.
This latter connection is now part of the main line between Sheffield and Leeds, and is used by CrossCountry services between Edinburgh Waverley and Birmingham New Street, and beyond. In addition, local trains on the Leeds – Rotherham Central – Sheffield route (Wakefield Line) use the spur and call at the station. This connecting line is now effectively the main line, and it splits from the S & K route immediately north of the station, with the track to York then bridging the GNR main line on its way northwards. This junction (and the adjacent loops to the south of the station) was controlled from the nearby Midland Railway signal box until May 2011, but following signalling equipment renewal work the area is now under the control of the ROC at York.
After the station was reduced to the status of an unstaffed halt in the 1980s, the station building was converted into a pub. This closed in the early 1990s and the building became derelict. However it has recently been restored by the town council, with funding from the Railway Heritage Trust, to include office space and a cafe.
A footbridge was opened at the station at the end of May 2010; previously passengers had to cross the tracks at a flat crossing.
Services
Monday to Saturdays, an hourly service operates from Moorthorpe to Leeds via Wakefield Westgate and to Sheffield. Sundays the service is two-hourly in both directions.
Three trains each way operate on the Dearne Valley line to York and Sheffield.
The station is less than away from South Elmsall railway station which means the South Kirkby/South Elmsall area has a half-hourly service (hourly Sundays) to Leeds.
The station is less than away from Frickley Athletic F.C.
Tickets
The station is in West Yorkshire but South Yorkshire PTE tickets are valid to and from this station on services into South Yorkshire.
References
External links
Category:Railway stations in Wakefield
Category:Former Swinton and Knottingley Joint Railway stations
Category:Railway stations served by Northern (train operating company)
Category:Railway stations opened in 1879
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Comparison of the effects of diltiazem and aspirin on platelet aggregation in cats.
Platelet aggregation in response to collagen (1 or 3 micrograms/ml), arachidonic acid (10(-2) M), and adenosine diphosphate (ADP, 2 microM) was compared in healthy cats treated with diltiazem (approximately 2 mg/kg body weight, q 8 hrs for 10 doses), aspirin (approximately 21 mg/kg body weight [1 baby aspirin], q 72 hrs for three doses), or a combination of diltiazem and aspirin. Baseline values obtained prior to treatment served as controls. Addition of arachidonic acid to blood resulted in an impedance change (i.e., aggregation) with time in samples from the nontreated cats and the cats treated with diltiazem, but the addition had no effect in blood from cats treated with aspirin alone or with a combination of diltiazem and aspirin. Platelet aggregation in response to either concentration of collagen or to ADP was not altered by any treatment. Secretion of adenosine triphosphate (ATP) from the platelets was measured when the aggregating agent was 3 micrograms/ml collagen; secretion was not affected by any treatment.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Finnish:
Tilaa, kuten, ja napsauta kellokuvaketta ♥
Minttu suklaajälkiruokajuhlat!
Kotitekoinen kakku täällä :))
Minttu suklaakuplitee
Hindi:
कृपया सदस्यता लें, लाइक करें और घंटी आइकन ♥ पर क्लिक करें
मिंट चॉकलेट मिष्ठान दावत!
घर का बना केक :))
मिंट चॉकलेट बबल टी
Indonesian:
Silakan berlangganan, suka, dan klik ikon bel ♥
Pesta Dessert Cokelat Mint!
Kue buatan sendiri di sini :))
Teh gelembung coklat mint
Dutch:
Abonneer u, like en klik op het belpictogram ♥
Dessertfeest met muntchocolade!
Zelfgemaakte cake hier :))
Mint chocolade bubble tea
Filipino:
Mangyaring mag-subscribe, tulad ng, at i-click ang icon ng kampanilya ♥
Mint tsokolate dessert pista!
Homemade cake dito :))
Mint na tsokolate bubble tea
English:
Please subscribe, like, and click the bell icon♥
Mint chocolate dessert feast!
Homemade cake here :))
Mint chocolate bubble tea
Polish:
Zasubskrybuj, polub i kliknij ikonę dzwonka ♥
Uczta z miętowo-czekoladowym deserem!
Domowe ciasto tutaj :))
Miętowa czekolada bąbelkowa
Japanese:
チャンネル登録、いいね、ベルのアイコンをクリックしてください♥
ミントチョコレートデザートのごちそう！
自家製ケーキはこちら:)）
ミントチョコレートバブルティー
Turkish:
Lütfen abone olun, beğenin ve zil simgesini tıklayın ♥
Naneli çikolatalı tatlı ziyafeti!
Burada ev yapımı kek :))
Nane çikolatalı kabarcık çayı
Thai:
กรุณาสมัครกดไลค์และคลิกไอคอนรูปกระดิ่ง♥
งานฉลองขนมช็อกโกแลตมิ้นต์!
เค้กโฮมเมดที่นี่ :))
ชาฟองช็อกโกแลตมิ้นต์
Modern Greek (1453-):
Εγγραφείτε, αρέσει και κάντε κλικ στο εικονίδιο καμπάνας ♥
Γιορτή επιδόρπιο με μέντα σοκολάτας!
Σπιτικό κέικ εδώ :))
Τσάι φυσαλίδας μέντα σοκολάτας
Bulgarian:
Моля, абонирайте се, харесайте и щракнете върху иконата на звънеца ♥
Десерт с ментов шоколадов десерт!
Домашна торта тук :))
Чай с ментов шоколад
Serbian:
Претплатите се, као, и кликните икону звона ♥
Благдански десерт са чоколадом!
Овде домаћа торта :))
Чај од метвице од чоколаде
Malay (macrolanguage):
Sila langgan, suka, dan klik ikon loceng ♥
Pesta pencuci mulut coklat mint!
Kek buatan sendiri di sini :))
Teh gelembung coklat pudina
Mongolian:
Та захиалга хийж like дарж үүлний дүрсийг ♥ дарна уу
Mint шоколадтай амттан найр!
Гэрийн бялуу энд байна :))
Mint шоколадтай хөөсний цай
Korean:
구독, 좋아요, 알림을 설정해주세요♥
민트 디저트 파티~!
제가 직접 구운 케이크도 있어요^0^
민트 초코 버블티
French:
Veuillez vous abonner, aimer et cliquer sur l'icône en forme de cloche ♥
Festin de dessert au chocolat à la menthe!
Gâteau fait maison ici :))
Bubble Tea au chocolat à la menthe
Vietnamese:
Hãy đăng ký, thích và nhấp vào biểu tượng chuông ♥
Bữa tiệc tráng miệng sô cô la bạc hà!
Bánh nhà làm đây :))
Trà trân châu sô cô la bạc hà
Italian:
Iscriviti, metti Mi piace e fai clic sull'icona della campana ♥
Festa del dessert al cioccolato alla menta!
Torta fatta in casa qui :))
Bubble tea alla menta e cioccolato
Portuguese:
Assine, curta e clique no ícone de sino ♥
Festa da sobremesa de chocolate com menta!
Bolo caseiro aqui :))
Chá bolha de chocolate com menta
Russian:
Подпишитесь, поставьте лайк и нажмите на значок колокольчика ♥
Мятно-шоколадный десертный праздник!
Домашний торт здесь :))
Мятный шоколадный пузырьковый чай
Tamil:
தயவுசெய்து குழுசேரவும், விரும்பவும், மணி ஐகானைக் கிளிக் செய்யவும்
புதினா சாக்லேட் இனிப்பு விருந்து!
வீட்டில் கேக் இங்கே :))
புதினா சாக்லேட் குமிழி தேநீர்
Kazakh:
Жазылыңыз, лайк басып, ♥ белгісін басыңыз
Жалбыз шоколады десерт мерекесі!
Үйдегі торт мына жерде :))
Жалбыз шоколады көпіршігі
Nepali (macrolanguage):
कृपया सदस्यता लिनुहोस्, जस्तै, र घण्टी प्रतिमा icon मा क्लिक गर्नुहोस्
पुदीना चकलेट डेजर्ट भोज!
यहाँ घर बनाइएको केक :))
टकसाल चकलेट बबल चिया
Spanish:
Suscríbete, dale me gusta y haz clic en el ícono de campana ♥
¡Fiesta de postre de chocolate con menta!
Pastel casero aquí :))
Té de burbujas de chocolate y menta
Swedish:
Prenumerera, gilla och klicka på klockikonen ♥
Mint chokladefterrätt fest!
Hemlagad tårta här :))
Mint choklad bubbla te
German:
Bitte abonnieren Sie und klicken Sie auf das Glockensymbol ♥
Minzschokoladen-Dessertfest!
Hausgemachter Kuchen hier :))
Minzschokoladen-Bubble-Tee
Arabic:
يرجى الاشتراك وإعجاب والنقر على أيقونة الجرس ♥
وليمة حلوى الشوكولاتة بالنعناع!
كعكة محلية الصنع هنا :))
شاي فقاعات النعناع بالشوكولاتة
Kazakh:
Уау өте жақсы !!
Қап! хаха
YUMMY ♥
Бұл менің талғамыма сәйкес келеді
Жалбыз саңырауқұлақ тағамдары
Turkish:
Oldukça iyi!!
Hata! haha
NEFİS ♥
Zevkime uyuyor
Nane mantarlı atıştırmalık
French:
Wow si bon!!
Oups! haha
YUMMY ♥
Ça correspond à mon goût
Collation aux champignons à la menthe
German:
Wow, so gut!!
Hoppla! Haha
YUMMY ♥
Es passt zu meinem Geschmack
Minzpilz-Snack
Arabic:
واو هذا جيد!!
وجه الفتاة! هاها
لذيذ ♥
يناسب ذوقي
وجبة خفيفة بالفطر بالنعناع
Russian:
Вау так хорошо !!
К сожалению! ха-ха
YUMMY ♥
Это соответствует моему вкусу
Закуска из мятных грибов
Serbian:
Вау тако добро !!
Упс! хаха
ИУММИ ♥
Одговара мом укусу
Залогај од менте са гљивама
Malay (macrolanguage):
Wah sangat bagus !!
Alamak! haha
YUMMY ♥
Ia menepati citarasa saya
Makanan ringan cendawan pudina
Polish:
Wow, tak dobrze !!
Ups! ha ha
YUMMY ♥
Pasuje do mojego gustu
Miętowa przekąska grzybowa
Spanish:
¡¡Qué bueno !!
¡Ups! jaja
Delicioso ♥
Se ajusta a mi gusto
Bocadillo de champiñones y menta
Portuguese:
Nossa tão bom!!
Ops! haha
YUMMY ♥
Se encaixa no meu gosto
Lanche de cogumelos com hortelã
Swedish:
Wow så bra !!
hoppsan! haha
YUMMY ♥
Det passar min smak
Mint svamp mellanmål
Modern Greek (1453-):
Ουάου τόσο καλό !!
Ωχ! χαχα
YUMMY ♥
Ταιριάζει στο γούστο μου
Μανιτάρι σνακ
Filipino:
Sobrang ganda !!
Oops! haha
YUMMY ♥
Naaangkop sa aking panlasa
Mack na meryenda ng kabute
Korean:
오 민트맛이에요~!ㅎㅎㅎㅎ
깜짝 놀랐어욯ㅎㅎㅎㅎㅎ
민트를 좋아하신다면 만족하실 맛이에요
제 입맛엔 딱이네요
민트 초코송이
Nepali (macrolanguage):
वाह कस्तो राम्रो !!
उफ्! हाह
YUMMY ♥
यो मेरो स्वाद फिट छ
पुदीना च्याउ खाजा
Vietnamese:
Wow quá tốt !!
Giáo sư! haha
YUMMY ♥
Nó phù hợp với sở thích của tôi
Snack nấm bạc hà
English:
Wow so good!!
Oops! haha
YUMMY♥
It fits my taste
Mint mushroom snack
Bulgarian:
Уау толкова добър !!
Ами сега! хаха
вкусен ♥
Подхожда на моя вкус
Закуска с мента с гъби
Tamil:
ஆஹா மிகவும் நல்லது !!
அச்சச்சோ! haha
யம்மி
இது என் சுவைக்கு பொருந்துகிறது
புதினா காளான் சிற்றுண்டி
Hindi:
वाह बहुत अच्छा!!
ऊप्स! haha
स्वादिष्ट ♥
यह मेरे स्वाद में फिट बैठता है
पुदीना मशरूम स्नैक
Dutch:
Wow zo goed!!
Oeps! haha
YUMMY ♥
Het past bij mijn smaak
Snack met munt en paddenstoelen
Finnish:
Vau niin hyvä!
Oho! haha
YUMMY ♥
Se sopii minun makuuni
Minttu sienipala
Japanese:
すごい！
おっとっと！ははは
YUMMY♥
それは私の好みに合います
ミントキノコスナック
Indonesian:
Wow bagus sekali !!
Ups! ha ha
YUMMY ♥
Ini sesuai dengan selera saya
Camilan jamur mint
Italian:
Wow così buono !!
Oops! haha
YUMMY ♥
Si adatta ai miei gusti
Spuntino ai funghi alla menta
Thai:
ว้าวดีมาก !!
อ๊ะ! ฮ่าฮ่า
YUMMY ♥
มันเข้ากับรสนิยมของฉัน
ขนมเห็ดมิ้นท์
Mongolian:
Хөөх маш сайн байна !!
Өө, хаха
YUMMY ♥
Энэ нь миний амтанд нийцдэг
Mint мөөгний зууш
Mongolian:
Мммм тийм амттай
Mine Oreo жигнэмэг
Malay (macrolanguage):
Mmmm sedap
Kuki Mint Oreo
German:
Mmmm so lecker
Minze Oreo Keks
Spanish:
Mmmm tan delicioso
Galleta Oreo de menta
Filipino:
Masarap si Mmmm
Cookie ng Mint Oreo
Russian:
Мммм так вкусно
Печенье Mint Oreo
Indonesian:
Mmmm enak sekali
Biskuit mint Oreo
Tamil:
ம்ம்ம்ம் மிகவும் சுவையாக இருக்கிறது
புதினா ஓரியோ குக்கீ
Kazakh:
Ммм өте керемет
Орео жалбызынан жасалған куки
Dutch:
Mmmm zo lekker
Mint Oreo-koekje
Modern Greek (1453-):
Μμμ τόσο νόστιμο
Μπισκότο μέντας Oreo
Turkish:
Mmmm çok lezzetli
Nane Oreo kurabiyesi
Finnish:
Mmmm niin herkullista
Mint Oreo -eväste
Thai:
อืมอร่อยมาก
คุกกี้ Mint Oreo
Nepali (macrolanguage):
Mmmm धेरै स्वादिष्ट
टकसाल Oreo कुकी
Polish:
Mmmm taki pyszny
Ciasteczko Mint Oreo
Arabic:
مممم لذيذ جدا
بسكويت النعناع أوريو
English:
Mmmm so delicious
Mint Oreo cookie
Bulgarian:
Mmmm толкова вкусно
Mint Oreo бисквитка
Portuguese:
Mmmm tão delicioso
Biscoito de menta Oreo
Hindi:
मम्मे इतने स्वादिष्ट
मिंट ओरेओ कुकी
Serbian:
Мммм тако укусно
Колачић Минт Орео
Vietnamese:
Mmmm rất ngon
Bánh quy Oreo bạc hà
Swedish:
Mmmm så läcker
Mint Oreo-cookie
Korean:
음~ 역시 맛있네용
민트 오레오
French:
Mmmm si délicieux
Biscuit Oreo à la menthe
Japanese:
うーんとてもおいしい
ミントオレオクッキー
Italian:
Mmmm così delizioso
Biscotto Menta Oreo
Turkish:
Renk çok güzel
ve tadı harika :))
Nane makaron
Aman Tanrım ♥♥
Nane çok derin
Vietnamese:
Màu sắc rất tốt
và hương vị tuyệt vời :))
Bánh macaron bạc hà
Omg tôi yêu nó ♥♥
Bạc hà rất sâu
Finnish:
Väri on niin hyvä
ja maistuu mahtavalta :))
Minttu makaroni
Omg rakastan sitä ♥♥
Minttu on niin syvä
Nepali (macrolanguage):
रंग यति राम्रो छ
र स्वाद अद्भुत :))
टकसाल म्याकारोन
Omg मलाई यो मन पर्छ ♥♥
टकसाल यति गहिरो छ
Portuguese:
A cor é tão boa
e tem um gosto incrível :))
Macaron de Menta
Nossa, eu amo isso ♥♥
A hortelã é tão profunda
English:
The color is so good
and tastes awesome :))
Mint macaron
Omg I love it♥♥
The mint is so deep
Thai:
สีน่ากินจัง
และรสชาติดีมาก :))
มาการองมิ้นท์
โอ้ฉันรักมัน♥♥
มิ้นท์ซึ้งมาก
German:
Die Farbe ist so gut
und schmeckt super :))
Minze Macaron
Omg ich liebe es ♥♥
Die Minze ist so tief
French:
La couleur est si bonne
et a un goût génial :))
Macaron à la menthe
Omg je l'aime ♥♥
La menthe est si profonde
Swedish:
Färgen är så bra
och smakar fantastiskt :))
Mint macaron
Omg jag älskar det ♥♥
Minten är så djup
Kazakh:
Түсі өте жақсы
және керемет талғаммен :))
Жалбыз макарон
Omg Мен оны жақсы көремін ♥♥
Жалбыз өте терең
Modern Greek (1453-):
Το χρώμα είναι τόσο καλό
και έχει υπέροχη γεύση :))
Μέντα macaron
Ωμ το λατρεύω ♥♥
Η μέντα είναι τόσο βαθιά
Bulgarian:
Цветът е толкова добър
и вкус е страхотен :))
Мента макарон
Omg, обичам го ♥♥
Ментата е толкова дълбока
Polish:
Kolor jest taki dobry
i smakuje niesamowicie :))
Macaron miętowy
Omg kocham to ♥♥
Mięta jest taka głęboka
Spanish:
El color es tan bueno
y sabe increíble :))
Macaron de menta
Dios mío, me encanta ♥♥
La menta es tan profunda
Russian:
Цвет такой хороший
и на вкус потрясающий :))
Мятный макарон
Боже, мне это нравится ♥♥
Мята такая глубокая
Dutch:
De kleur is zo goed
en smaakt geweldig :))
Munt macaron
Omg ik vind het geweldig ♥♥
De munt is zo diep
Filipino:
Napakaganda ng kulay
at panlasa kahanga-hangang :))
Mint macaron
Omg mahal ko ito ♥♥
Malalim ang mint
Arabic:
اللون جيد جدا
وطعمه رائع :))
النعناع معكرون
يا إلهي أنا أحبه ♥♥
النعناع عميق جدا
Indonesian:
Warnanya sangat bagus
dan rasanya luar biasa :))
Macaron mint
Ya ampun, aku menyukainya ♥♥
Permennya sangat dalam
Malay (macrolanguage):
Warnanya sangat bagus
dan rasanya hebat :))
Mint macaron
Omg saya suka ♥♥
Pudina begitu dalam
Japanese:
色がとても良いです
と素晴らしい味:)）
ミントマカロン
Omg I love it♥♥
ミントはとても深い
Tamil:
நிறம் மிகவும் நன்றாக இருக்கிறது
மற்றும் சுவை அற்புதம் :))
புதினா மாக்கரோன்
ஓம் நான் அதை விரும்புகிறேன்
புதினா மிகவும் ஆழமானது
Serbian:
Боја је тако добра
и укусно је :))
Мента мацарон
Омг, волим то ♥♥
Ковница је толико дубока
Hindi:
रंग इतना अच्छा है
और कमाल का स्वाद :))
मिंट मैकरॉन
ओमजी मैं इसे प्यार करता हूं ♥♥
टकसाल इतनी गहरी है
Korean:
색이 정말 예쁘죠?
맛도 아주 좋아요^0^
민트 마카롱
이거 너무너무 맛있어요♥
아주 진한 민트 초코 맛이에요
Mongolian:
Өнгө нь маш сайн
их сайхан амттай
Mint macaron
Omg би үүнд хайртай ♥♥
Гаа нь маш гүнзгий
Italian:
Il colore è così buono
e ha un sapore fantastico :))
Macaron alla menta
Omg lo adoro ♥♥
La menta è così profonda
Spanish:
Chocolate
¡¡¡Lo hice yo!!! jaja
Es dulce pero no demasiado :)
Indonesian:
Cokelat
Saya membuatnya sendiri!!! ha ha
Ini manis tapi tidak terlalu banyak :)
Russian:
Шоколад
Я сам это сделал!!! ха-ха
Сладко, но не слишком :)
Thai:
ช็อคโกแลต
ฉันทำเอง !!! ฮ่าฮ่า
มันหวาน แต่ไม่มาก :)
Tamil:
சாக்லேட்
நானே செய்தேன் !!! haha
இது இனிமையானது ஆனால் அதிகமாக இல்லை :)
Hindi:
चॉकलेट
ये मैंने स्वयं बनाया था!!! haha
यह मीठा है लेकिन बहुत ज्यादा नहीं है :)
German:
Schokolade
Ich habe es selbst gemacht!!! Haha
Es ist süß, aber nicht zu viel :)
Modern Greek (1453-):
Σοκολάτα
Το έφτιαξα μόνος μου!!! χαχα
Είναι γλυκό αλλά όχι πάρα πολύ :)
Serbian:
Чоколада
Ја сам то направио!!! хаха
Слатко је, али не превише :)
Italian:
Cioccolato
L'ho fatto da solo!!! haha
È dolce ma non troppo :)
Nepali (macrolanguage):
चकलेट
मैले यो आफैं बनाएँ !!! हाह
यो मिठो छ तर धेरै छैन :)
Malay (macrolanguage):
Coklat
Saya membuatnya sendiri !!! haha
Ia manis tetapi tidak terlalu banyak :)
Finnish:
Suklaa
Tein sen itse!!! haha
Se on suloista, mutta ei liikaa :)
Filipino:
Tsokolate
Ginawa ko ito !!! haha
Ito ay matamis ngunit hindi masyadong maraming :)
Arabic:
شوكولاتة
صنعته بنفسي!!! هاها
إنه حلو لكن ليس كثيرًا :)
Korean:
초콜릿
케이크가 너무 잘 구워졌어요
많이 달지 않고 아주 맛있어요
Swedish:
Choklad
Jag gjorde det själv!!! haha
Det är söt men inte för mycket :)
French:
Chocolat
Je l'ai fait moi-même!!! haha
C'est doux mais pas trop :)
Turkish:
Çikolata
Onu kendim yaptım!!! haha
Tatlı ama çok fazla değil :)
Vietnamese:
Sô cô la
Tôi đã tự làm nó!!! haha
Nó ngọt ngào nhưng không quá nhiều :)
Polish:
Czekolada
Sam to zrobiłem!!! ha ha
Jest słodko, ale nie za dużo :)
Kazakh:
Шоколад
Мен оны өзім жасадым !!! хаха
Бұл тәтті, бірақ көп емес :)
Japanese:
チョコレート
自分で作ったよ!!!ははは
それは甘いですが、多すぎません:)
English:
Chocolate
I made it myself!!! haha
It's sweet but not too much :)
Dutch:
Chocola
Ik heb het zelf gemaakt!!! haha
Het is lief maar niet te veel :)
Portuguese:
Chocolate
Eu fiz isso sozinho!!! haha
É doce, mas não muito :)
Bulgarian:
Шоколад
Направих го сам !!! хаха
Сладко е, но не много :)
Mongolian:
Шоколад
Би өөрөө хийсэн шүү !!! хаха
Энэ нь чихэрлэг гэхдээ хэт их биш шүү :)
Modern Greek (1453-):
Το καλύτερο κέικ που είχα ποτέ!
Θα γυρίσω σύντομα το βίντεο συνταγών
Μεταφόρτωσα τη συνταγή Croffle στο υπο κανάλι μου
Εάν δεν το έχετε δει ακόμα, ελέγξτε το τώρα!
Finnish:
Paras kakku, mitä minulla on koskaan ollut!
Kuvaan reseptivideon pian
Latasin Crofflen reseptin alakanavalleni
Jos et ole vielä nähnyt sitä, tarkista se heti!
Russian:
Лучший торт, который у меня когда-либо был!
Я скоро сниму видео с рецептом
Я загрузил рецепт Croffle на свой подканал
Если вы еще не видели его, посмотрите прямо сейчас!
Spanish:
¡El mejor pastel que he comido!
Pronto filmaré el video de la receta
Subí la receta de Croffle en mi subcanal
Si aún no lo has visto, ¡compruébalo ahora mismo!
Japanese:
私が今まで食べた中で最高のケーキ！
すぐにレシピ動画を撮ります
サブチャンネルにCroffleレシピをアップロードしました
まだご覧になっていない場合は、今すぐチェックしてください！
Kazakh:
Менде болған ең жақсы торт!
Мен жақында рецепт бейнесін түсіремін
Мен Croffle рецептін ішкі арнаға жүктедім
Егер сіз оны әлі көрмесеңіз, дәл қазір тексеріңіз!
English:
The best cake I've ever had!
I'll film the recipe video soon
I uploaded Croffle recipe on my sub channel
If you haven't seen it yet, check it right now!
Portuguese:
O melhor bolo que já comi!
Vou filmar o vídeo da receita em breve
Carreguei a receita do Croffle no meu subcanal
Se você ainda não viu, verifique agora mesmo!
Dutch:
De beste cake die ik ooit heb gehad!
Ik film binnenkort de receptvideo
Ik heb het Croffle-recept geüpload op mijn subkanaal
Als je het nog niet hebt gezien, controleer het dan nu!
French:
Le meilleur gâteau que j'ai jamais eu!
Je vais bientôt filmer la vidéo de la recette
J'ai téléchargé la recette Croffle sur ma sous-chaîne
Si vous ne l'avez pas encore vu, vérifiez-le maintenant!
Mongolian:
Миний хийсэн хамгийн сайхан бялуу!
Удахгүй жор жороор видеог хийх болно
Би Croffle жорыг өөрийн дэд суваг дээрээ байршуулсан
Хэрэв та хараахан харж амжаагүй бол яг одоо шалгаарай!
Polish:
Najlepsze ciasto, jakie kiedykolwiek jadłem!
Wkrótce nakręcę film z przepisem
Wgrałem przepis na Croffle na moim kanale podrzędnym
Jeśli jeszcze tego nie widziałeś, sprawdź to teraz!
German:
Der beste Kuchen, den ich je hatte!
Ich werde das Rezeptvideo bald filmen
Ich habe das Croffle-Rezept auf meinen Subkanal hochgeladen
Wenn Sie es noch nicht gesehen haben, überprüfen Sie es jetzt!
Turkish:
Şimdiye kadar yaşadığınız en iyi pasta!
Tarif videosunu yakında çekeceğim
Croffle tarifini alt kanalıma yükledim
Henüz görmediyseniz, hemen kontrol edin!
Serbian:
Најбоља торта коју сам икад имао!
Ускоро ћу снимити видео запис о рецепту
Отпремио сам Цроффле рецепт на свој суб канал
Ако га још нисте видели, проверите одмах!
Vietnamese:
Chiếc bánh ngon nhất mà tôi từng có!
Tôi sẽ quay video công thức sớm
Tôi đã tải lên công thức Croffle trên kênh phụ của mình
Nếu bạn chưa thấy nó, hãy kiểm tra nó ngay bây giờ!
Swedish:
Den bästa tårta som jag någonsin har haft!
Jag ska filma receptvideoen snart
Jag laddade upp Croffle-recept på min underkanal
Om du inte har sett det ännu, kolla det nu!
Tamil:
நான் வைத்திருந்த சிறந்த கேக்!
ரெசிபி வீடியோவை விரைவில் படமாக்குவேன்
எனது துணை சேனலில் கிராஃபில் செய்முறையை பதிவேற்றினேன்
நீங்கள் இதுவரை பார்க்கவில்லை என்றால், இப்போதே சரிபார்க்கவும்!
Arabic:
أفضل كعكة تناولتها على الإطلاق!
سأقوم بتصوير فيديو الوصفة قريبًا
لقد قمت بتحميل وصفة Croffle على قناتي الفرعية
إذا لم تره بعد ، فتحقق منه الآن!
Filipino:
Ang pinakamagandang cake na mayroon ako!
Papasok ko agad ang video ng recipe
Nag-upload ako ng Croffle recipe sa aking sub channel
Kung hindi mo pa ito nakita, suriin ito ngayon!
Malay (macrolanguage):
Kek terbaik yang pernah saya ada!
Saya akan merakam video resipi tidak lama lagi
Saya memuat naik resipi Croffle di sub saluran saya
Sekiranya anda belum melihatnya, periksa sekarang juga!
Indonesian:
Kue terbaik yang pernah kumiliki!
Saya akan segera merekam video resepnya
Saya mengupload resep Croffle di sub saluran saya
Jika Anda belum melihatnya, periksa sekarang juga!
Korean:
제가 만들었지만 정말 맛있네요 ㅎㅎㅎㅎㅎㅎ
나중에 레시피 영상도 찍어볼게요
며칠 전에 크로플 레시피를 두 번째 채널에 올렸어요
아직 못 보신 분들은 두 번째 채널을 확인해주세요~
Italian:
La torta migliore che abbia mai avuto!
Filmerò presto il video della ricetta
Ho caricato la ricetta di Croffle sul mio canale secondario
Se non l'hai ancora visto, controllalo subito!
Hindi:
अब तक का सबसे अच्छा केक!
मैं जल्द ही रेसिपी वीडियो फिल्माऊंगा
मैंने अपने उप चैनल पर क्रॉफ़ल रेसिपी अपलोड की
यदि आपने इसे अभी तक नहीं देखा है, तो इसे अभी जांचें!
Bulgarian:
Най-добрата торта, която съм имала!
Ще заснема видеото за рецептата скоро
Качих рецептата на Croffle в моя подканал
Ако още не сте го виждали, проверете веднага!
Thai:
เค้กที่ดีที่สุดที่ฉันเคยทานมา!
ฉันจะถ่ายทำวิดีโอสูตรอาหารเร็ว ๆ นี้
ฉันอัปโหลดสูตร Croffle ในช่องย่อยของฉัน
หากคุณยังไม่เคยเห็นมาตรวจสอบตอนนี้!
Nepali (macrolanguage):
मसँग भएको सब भन्दा राम्रो केक!
म चाँडै यो नुस्खा भिडियो फिल्म गर्नेछु
मैले मेरा सब च्यानलमा क्रफल रेसिपी अपलोड गरें
यदि तपाईंले अहिले सम्म यो देख्नु भएको छैन भने, अहिले नै जाँच गर्नुहोस्!
|
{
"pile_set_name": "YoutubeSubtitles"
}
|
162 F.3d 1153
Bruce L. Wishnefskyv.David W. Addy
NO. 97-1629
United States Court of Appeals,Third Circuit.
June 25, 1998
1
Appeal From: E.D.Pa. ,No.97cv02500 ,
969 F.Supp. 953
2
Affirmed.
|
{
"pile_set_name": "FreeLaw"
}
|
Aphonogelia
Aphonogelia is a rare neuropsychological condition with which a person cannot laugh audibly.
References
Category:Neuropsychology
Category:Laughter
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Q:
Use the increased navigation-bar title in iOS 11
iOS 11 Beta 1 uses the increased navigation-bar title for almost all system-apps (it started doing this in iOS 10 and the Music app). I am wondering if Apple has a public API for this coming in iOS 11, or whether it will stay private for now.
The behavior is that the title has an increased font-size, is left aligned and will move to the navigation-bar once the user scrolls down. I've attached some screens showing this behavior in the Messages app here:
Although I could not find any reference in the UINavigationController and UINavigationBar so far, maybe someone knows some more details!
A:
The only change done to UINavigationBar API for iOS 11 is prefersLargeTitles.
Documentation here: https://developer.apple.com/documentation/uikit/uinavigationbar/
You can do it to your own apps with one small change: check "Prefers Large Titles" for your navigation bar in IB, or if you prefer to do it in code using:
navigationController?.navigationBar.prefersLargeTitles = true
If you need to change the text attributes of the large title you need to use the new largeTitleTextAttributes property on UINavigationBar:
UINavigationBar.appearance().largeTitleTextAttributes = [
NSAttributedString.Key.foregroundColor: UIColor.black
]
A:
UINavigationBar has a prefersLargeTitles: Bool property. Docs here.
class UINavigationBar {
var prefersLargeTitles: Bool
}
UINavigationItem has a largeTitleDisplayMode: UINavigationItem.LargeTitleDisplayMode property. Docs here.
class UINavigationItem {
var largeTitleDisplayMode: LargeTitleDisplayMode
}
Both of these can be modified in the Interface Builder.
To turn on this behavior set navigationController.navigationBar.prefersLargeTitles to true. Then you can control each individual view controller in the navigation controller stack by setting navigationItem.largeTitleDisplayMode.
The general design guidelines by Apple are that large titles shouldn't be used everywhere (for example, the Clock app does not use them), and it's generally preferred that only the first level of the navigation controller uses the large titles. However, these are just general guidelines.
Large titles are introduced in What's New in Cocoa Touch video (7:37).
A:
Just check "Prefers Large Titles" in the Navigation Bar attribute inspector in Storyboard / Interface Builder:
|
{
"pile_set_name": "StackExchange"
}
|
On the night of the 1980 US presidential election, young David Barron dons a Ronald Reagan mask and brutally murders his ex-porn star mother, Scorchy McDaniels (Rae). Thirty-eight years later, and on the eve of the mid-term elections, David (Adams) escapes from the Lar-A-Mago sanitarium where he’s spent the intervening years. Returning to his home town of Libertybelle, David takes to wearing a Donald Trump mask and hanging out at his childhood home. Meanwhile, Dr Lutin (Sing), his doctor, heads there in the hope of finding David – though he has an ulterior motive for doing so. In the same neighbourhood, best friends Lana (Attaie), Blanca (Perez), and Medjine (Boone) are preparing to have a pre-election party ahead of their participation in an anti-Republican rally on the day. Along with Blanca’s younger brother, Pepe (Jorgensen), and their transgender friend, Gabriel (Sander), the party gets off to a good start, but it isn’t long before David is picking them off one by one, while the town sheriff (Rosete) does his best to come to their aid before it’s entirely too late…
The pitch must have been a fairly simple one: hey, why don’t we make a spoof of the original Halloween where instead of a Captain Kirk mask, the killer wears a Donald Trump mask instead? And the response must have been equally simple: great idea, go make it. But in the tradition of simple ideas made on a restricted budget, President Evil is an uneven, occasionally inspired, occasionally woeful movie with a ton of good intentions that don’t always pay off. It begins with an opening credits sequence that replicates the style of Halloween’s own opening credits, but replaces the jack o’ lantern with a Trump mask. Then there’s an updated recreation of the young Michael Myers’ murder of his sister that is shot entirely from David’s point of view and ends with him being unmasked outside his home. So far, so reassuringly competent homage, though with the kind of comedic elements that reveal the makers’ broader intentions for their story. Nods and winks in the direction of John Carpenter’s seminal movie follow, as well as Easter eggs that reference some of his other movies, while the script also adds further homages from the likes of Psycho (1960) and Young Frankenstein (1974).
The comedy is a mixed bag all by itself, and ranges from deft visual flourishes (David’s Trump mask hides someone who looks like Trump), to irritating bouts of frat humour (best summed up by Miles’ popping up at odd moments as characters as varied as a perverted priest and a Jared Kushner look-a-like), and further Mel Brooks’ appropriations (“Be a Smarty and Join the Republican Party”). Like Halloween, there’s a minimum of blood and gore, but there’s a singular lack of tension throughout, and the killings are often poorly staged and framed. The performances are broadly acceptable for this sort of thing, though Attaie does make for an appealing heroine, and Lowry seems more confident when bending the knee to Carpenter’s original than he does with the newer material; it’s as if the obvious difference between them was a given he had no control over. But if there’s one aspect that the script – by Lowry and Gregory P. Wolk – does get right, it’s in depicting the anger and distrust of ethnic minorities in current day America towards the xenophobic attitudes of the predominantly white, privileged political system. The movie is strident in its approach, but is also unapologetic about being so, and on that level – and like the best of horror movies – proves to be a telling reflection of a section of US society’s real fears.
Rating: 5/10 – though there’s much that doesn’t work, and much else that should have been jettisoned at the earliest opportunity, there’s still much to enjoy in President Evil, not the least of which is the way it lampoons Donald Trump and his ill-advised ramblings; to call this a post-millennial horror comedy for post-millennials who believe they might be the last generation able to appreciate something like this, may be stretching things, but when it’s en point, there’s nothing “Fake” about it.
|
{
"pile_set_name": "Pile-CC"
}
|
Evidence in a child pornography trial has been thrown out because the US government refuses to detail how it hacked the website allegedly visited by the defendant.
The FBI says the hacking method, referred to as a Network Investigation Technique, or NIT, allowed the bureau to track Jay Michaud after he visited a hidden website on the so-called Dark Web, leading to charges of possessing child pornography. Defense attorneys say a government explanation could show that the method yielded unreliable information.
The case is part of a growing debate over government hacking in criminal cases. Michaud is one of several people facing charges after the bureau used the hacking method to infiltrate the hidden child pornography website and identify the computers of those who visited it. A handful of those cases have seen evidence tossed out as well. In addition to concerns over the reliability of the hack raised by Michaud's attorneys, judges have raised concerns over warrants that allow the government to target a computer even when its location is unknown.
As the resulting cases have unfolded, the US Senate is considering a change to federal judicial rules that would let judges sign warrants to permit the government to target computers outside their jurisdiction when their locations are unknown.
In the case of Michaud, who was a middle school teacher in Vancouver, Wash., the technique led police to obtain a search warrant last year for his home, where they allegedly found a cell phone and two thumb drives containing child pornography. That evidence is no longer part of the case.
"Evidence of the NIT, the search warrant issued based on the NIT, and the fruits of that warrant should be excluded and should not be offered in evidence at trial," Judge Robert J. Bryan of the US District Court of the Western District of Washington wrote in his opinion Wednesday, issued after a hearing.
Michaud allegedly visited the child pornography site through the Tor browser, a tool that lets Internet users disguise their locations and visit hidden areas of the Web.
His attorneys argued that he had a right to know exactly how the government carried out the alleged hack, "given the sophistication of the FBI's surveillance technology and the evidence that it has misled the courts in other cases about that technology."
The brief cited a report from the Associated Press about cases in which child pornography was found stashed on the hacked computers of innocent people.
Colin Fieman, a federal public defender representing Michaud, did not respond to a request for comment.
The judge had already ordered the government to turn over its code, but the US Attorney asked him to reconsider. After pointing out the files containing alleged child pornography were found on the defendant's thumb drives and cell phone, the US Attorney's brief said, "any concern about corruption or other errors that might cast doubt on the accuracy of the information obtained through the NIT...can be addressed by review of the information that was actually collected."
The US Department of Justice did not respond to a request for comment on the case.
While Bryan ruled Wednesday to exclude evidence from the trial, he also wrote that the case shouldn't be dismissed.
|
{
"pile_set_name": "OpenWebText2"
}
|
Richard Lowther, 2nd Viscount Lonsdale
Richard Lowther, 2nd Viscount Lonsdale (1692 – 1 December 1713) was an English nobleman, the eldest son of John Lowther, 1st Viscount Lonsdale and Katherine Thynne.
He succeeded his father at the age of eight, but died in 1713 a few months after reaching his majority.
References
Category:1692 births
Category:1713 deaths
Richard
Category:Viscounts Lonsdale
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Lesson timetable
Learning outcomes
The course aims at providing the students with simple applied tools for the econometric analysis based on the use of basic notions of statistics and econometrics. The course is devoted to the analysis of both macro and micro data. The program includes the statistical analysis of economic data bases and exercises applied to economic cases of behavioral and policy relevance. The course is also offered as an e-learning course: http://pilar.univr.it/docebo.
|
{
"pile_set_name": "Pile-CC"
}
|
Government requests for user data from Google hit an all-time high in the second half of 2015, the internet company revealed on Monday.
Through July to December 2015, governments from around the globe made 40,677 requests, impacting as many as 81,311 user accounts. That's an 18 percent spike from the first half of 2015, when government requests for data impacted 68,908 users.
By far and away, the most requests came from the United States, which made 12,523 data requests for this reporting period. The requests impacted 27,157 users or accounts.
Google reports the number of user data requests it has received every six-month period going as far back as the second half of 2009. It started detailing the number of users and/or accounts impacted in the first half of 2011. "Usage of our services have increased every year, and so have the user data request numbers," the company noted.
Since the second half of 2010, Google has reported the percentage of user data requests it at least partially complies with. For the second half of 2015, the company produced at least some data for 64 percent of requests. That figure has been about the same since 2013, but it's been trending slightly downward. Google complied with 79 percent of requests from the United States.
For its reporting on US requests, Google details the specific type of criminal legal requests it received. For instance, "pen register" rules allow the US government to collect information like IP addresses or telephone numbers dialed, if the information likely to be obtained is relevant to national security or a criminal investigation. Google reports that it received 241 pen register orders for the second half of 2015, impacting 439 accounts. The company complied with 92 percent.
Google only counts criminal legal requests in its total tally for US data requests. However, the company also provides a range of how many requests it receives under two national security tools: The Foreign Intelligence Surveillance Act (FISA) and National Security Letters. This data is subject to a six-month reporting delay.
For the first half of 2015, Google reported that it once again received somewhere between zero and 499 non-content related FISA requests. It received between 500 and 999 content FISA requests, impacting between 16,000 and 16,499 users. That's a slight drop from the second half of 2014, when somewhere between 17,500 and 17,999 users were impacted.
Google reported receiving under 500 National Security Letters for the first half of 2015, impacting between 500 and 999 accounts.
In a blog post, Google said it is pleased with some developments in surveillance laws, including the adoption of the Privacy Shield, the new pact between the EU and the US that sets the terms for trans-Atlantic transfers of personal data. Google also commended the passage of the US Judicial Redress Act, which lets Europeans bring civil actions in the US if agencies there intentionally violate the US Privacy Act when handling personal data.
"The distinctions that U.S. privacy and surveillance laws make between U.S. and non-U.S. persons are increasingly obsolete in a world where communications primarily take place over a global medium: the Internet," wrote Google legal director Richard Salgado.
|
{
"pile_set_name": "OpenWebText2"
}
|
Differential mRNA expression of TACR1 after ischemic peritoneal trauma: a pilot animal study.
Experimental trial in an in vivo animal model in the laboratory facilities of a university department of obstetrics and gynecology and a microarray facility using seventeen female Wistar rats to investigate the regional expression level of TACR1 at specific locations in the peritoneum in a rodent animal model of post-operative adhesions. Peritoneal adhesions were induced by the placement of three unilateral ischemic lesions. A time course experiment was performed to identify when adhesions form in this model to determine the optimal time for tissue harvesting. To this effect, second look analysis for adhesion scoring occurred after day 1, 3 and 5. Eighteen tissue samples from the adhesiogenic lesions and the contralateral non-adhesiogenic peritoneum were harvested from n = 3 animals at day 3 for quantitative real-time PCR analysis. After 1 day, no adhesions were macroscopically detectable. After 3 days, adhesions were detectable which could be separated easily by gravity. After 5 days, all animals had formed adhesions and strong traction was required for adhesiolysis. The adhesions always formed to the ischemic part of the lesions. Quantitative PCR analysis after 3 days demonstrated down-regulation of TACR1 mRNA in the adhesiogenic peritoneum of the lesions compared to non-adhesiogenic peritoneum on the contralateral side. This difference was statistically highly significant (p < 0.01). In the ischemic lesion model of adhesiogenesis, TACR1 is differentially expressed between adhesiogenic peritoneum and non-adhesiogenic peritoneum at the time-point of adhesion formation.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Dollar Falls Sharply
On Tame Inflation
March Data Signal
Fed May Cut Rates;
Pound Gets Big Lift
The dollar hit a 15-year low against sterling and neared an all-time low versus the euro after tame U.S. inflation data increased the odds the Federal Reserve may cut interest rates in coming months.
The economic data pushed the euro to an intraday peak of $1.3595, its highest point since December 2004 and close to its all-time high of $1.3670. Sterling, which earlier in the session climbed above $2 for the first time since 1992, pushed even higher on the inflation data, topping out at $2.0075.
|
{
"pile_set_name": "Pile-CC"
}
|
My drug of choice (and why it’s OK).
The other day I sat down in front of the computer to do something of vital importance (that I can’t remember) like pay a bill or register one of the Things in a fall activity or check out the fall boots at Zappos, and 30 (75) minutes later had produced this –
It’s the DVR schedule I made for the new fall season trying to see if the new shows I want to record will conflict with our old favorites…of which we may have just a few. (Turns out Tuesday night is gonna be pretty rough on the DVR. Too many shows on too many channels. I now have made a chart that rivals a Final Four bracket with which DVRs in this house will be recording what, and what shows we’ll have to catch up on with OnDemand. See? I’m totally good for more than scrubbing toilets and taxi-service.) And if you’re wondering, the highlighted shows are ones we cannot miss. The favorites of the favorites. I know X-Factor hasn’t started yet, but we can’t wait. It’ll be like driving past the accident that you cannot take your eyes away from and then realizing that Britney freakin’ Spears was the driver.
Looking at the above photo now, I realize I should be embarrassed about our viewing habits (and perhaps choices), but I’ve long since stopped apologizing for the shallow depth of my interests. Whatever. We love TV. And it’s OK.
And what got me distracted from my Zappos shopping bill paying in the first place was a promo I watched for a new show on ABC called “The New Normal” which is created by Ryan Murphy, who also created “GLEE” (which Thing 1 and I never miss) and “American Horror Story” (which Husband never misses, but what he hasto watch alone and with me 2 floors up) and which is creating quite a stir already.
~ Apparently, some people find this show about a few people making decisions based on love and commitment and generosity offensive, but yet the Kardashians and the Jersey Shore idiots (who are such genuine contributing members of society) are acceptable and allowed to air in all 50 states. ?? Confuses me. And I can’t wait to watch.
So anyway, that got me started searching for clips of other new shows slated for the fall season, and then it was all over. I started YouTubing every trailer for every new show listed (you can see how this took awhile…and how full my day is…and why the Things do not have any clean socks) and getting ridiculously excited for the new crop of shows. And then I remembered this.
Do you remember ‘The Playboy Club’?
No?
That’s because it ran for 3 weeks.
After a full summer of hype, I just knew it was gonna be my new go-to show. The one I couldn’t miss. I watched one awful episode. Curse you ABC and Eddie Cibrian for getting me all worked up and leaving me as disappointed as a girl who doesn’t get kissed after prom (but I wouldn’t know anything about that…).
It’s risky, dating a new show. Especially if it’s up against one of your old favorites. And committing takes literally half a season of episodes. Some shows hook and sink you right off the bat (‘Revenge’) and some are slow to warm up (‘New Girl’ certainly didn’t have us at “hello” but those kids definitely grew on us) and some shows are so awful right out of the gate you wonder how they’re still on the air (ahem…Two Broke Girls…cough…cough…).
The new TV season is a lot like the new NFL season, if you think about it (which I am because Husband just planted the seed in my head.) In the beginning there’s such hope! Optimism that you’re gonna capture that magic of the first season of FRIENDS or the 2009 MN Vikings’ season (at least the first part).
You look at the new fall TV lineup the same way that you do the new season’s draft class of (insert your favorite team here).
And maybe the first few episodes are marginal at best, but like pre-season games, you stick with it and keep your fingers crossed that they’ll pull it together.
And then, more often than not, like in the case of ‘Up All Night’ or the 2011 MN Vikings’ season, you reach the end of November switching channels….and teams.
It’s a gamble, but we roll the dice every fall.
What will be the new hits this year? Any guesses? There’s a bunch of new dramas with underlying themes of horror on the roster (thanks to the success of ‘American Horror Story’), and a bunch of new sitcoms with new parents and their adorable babies and friends who hang out in bars (thanks to…oh, yeah…every other sitcom that’s ever been written) and a bunch of shows on non-network channels that will be pushing the ratings limit and basically be about nothing other than sex and drinking and shocking language (HBO’s ‘Girls’ can’t return fast enough for me).
When the new season starts next week, we’ll be ready (thanks to my handy list. Btw – you’re welcome, Husband). And it’s OK that we spend waste literally hours of our lives sitting in our chairs watching TV. It’s our quality time. Our ‘we’ time. We have wine. We eat pistachios. We I hit ‘pause’ on the remote occasionally every few minutes to ask a question or make an observation or tell Husband a story that the plot of the show reminded me of (that no, cannot wait). From ‘Friends’ to ‘Mad About You’ to ‘ER’ to ‘Coach’ to ‘Ellen’ (nope, not the talk show – the sitcom. Yeah, we’re that old) – the list of shows we’ve laughed at and cried with over the past 20 years have been some of our favorite ways to spend time together.
So I won’t apologize for being addicted to television.
Oh, but the drug of choice that I’m talking about in the title up there?
Share this:
Sarah Grace-that schedule is quite impressive! and i know i’m about to lose some cool points, but i really liked the playboy club! and pan-am, which got canceled too. o well,,,good thing i have a bunch of other new shows on my list to watch (thanks to your schedule!)
Carolyn-OMG! I totally need to work on my DVR list. You’re a pro! Jake and I actually had to get a second DVR because we fought over what to record too much. LOL I am excited for Nashville… looks good! 🙂ReplyCancel
JanM ♥-Oh, I’m the same way! I’m scared of Thursdays because from 8-9 there will be Vampire Diaries and from 9-10 Grey’s Anatomy. Looks like I need to make dinner for two nights on Wednesdays 😉
|
{
"pile_set_name": "Pile-CC"
}
|
The capital awoke to thick fog this morning, and the moment it lifted has been captured in a dramatic timelapse video.
Filmed from The View from the Shard, the observation deck on the 69th and 72nd floors of the UK’s tallest building, the footage shows the cloud moving to reveal London landmarks such as the Gherkin, the Cheesegrater and the Walkie-Talkie (aka 30 St Mary Axe, the Leadenhall Building and 20 Fenchurch Street respectively), as well as Tower Bridge and the Tower of London.
In the first shot the Cheesegrater’s red lights can be seen shining through the fog until the sun rises, and the rest of the city begins to emerge. The second shot captures a number of boats travelling along the river, as well as trains coming into and out of London Bridge station.
With 95 storeys, the Shard stands 1,016ft high; the observation deck is around 800ft up.
|
{
"pile_set_name": "OpenWebText2"
}
|
QQ Group
QQ Group
Tag: slot scatter symbols
Slot mobile machine gaming has been so popular nowadays. As it has been classified to be the most loved casino games. Because of the fact, that slot machine gaming is so easy to understand and very simple to play. That is why many other casino fans, had switched from being a fan of other casino game to be a solid fan of a slot mobile machine game because of Slot mobile show amazing scatter symbol that generates wonderful lucky wins. Well, this has been a very undeniable event, because it has been the truth, that slot machine games can really attract more and more players, because of its incredibly unique features like can generates wonderful lucky wins. Such as the appearance of a scatter symbol, which gives a different kind of thrill.
Before playing a slot mobile machine game, it is must to fully understand first the concept of the game, to be it appreciated so much. Because having no idea on how this game works, will be a very big problem. Because such amount of money and effort will be wasted. And to resolve such case, it is better to do research on how this wonderful game really works, to even have a possibility of winning.
After knowing how the slot machine game works, it is now a must to understand the other features of slot machine games. Such as the so-called scatter symbol and lucky wins. What does a scatter symbol do really means? It is just basically a key to unlocking various fun bonus features designed within a slot mobile game, because of its unique characteristics it is even called as the slot player’s best friend, as it will be a great help on the entire slot machine gaming.
Generates a wonderful lucky winning
A slot machine game surely offers its players to generates wonderful lucky wins, with different procedures. And such procedure like the use of a scatter symbol. It is really beneficial to use a scatter symbol as it just has to appear on the reels to open up free spins, mini-games, or other bonuses. Because of such features, scatter symbols are considered to be typically one of the highest paying symbols in a slot mobile game and are represented by a specific graphic that is unique to the theme of the particular slot title. Making it more attracting to use.
So how does a scatter symbol generates? Simply, because it requires the certainful number of scatter symbols to appear on a single spin to unlock a feature. Such details are based on the game’s pay table. Once the needed number of wonderful scatter symbols appears on a single spin, the game will introduce its promised lucky wins feature. Usually, free spins are presented, but sometimes the scatter symbol is used to open up other interactive bonus games that will surely help a player to win bigger.
A wonderful scatter symbol is recognized to be a game specific one. As it generates a bit differently across the board, according to the slot mobile played by a gamer. It is also presumed that scatter icons and may sometimes just payout on a winning combination, making it a little bit tricky to use.
No game can’t be played nowadays with using our heads. We must learn that once the opportunity knocks on our door, we better grab it or regret ignoring it. Just like Slot mobile show amazing scatter symbol that generates wonderful lucky wins, use it wisely or else regret for nothing.
|
{
"pile_set_name": "Pile-CC"
}
|
John Thomas Lang was born on 21 December 1876 on George Street, Sydney, close to the present site of The Metro Theatre (between Bathurst Street and Liverpool Street). He was the third son (and sixth of ten children) of James Henry Lang, a watchmaker born in Edinburgh, Scotland, and Mary Whelan, a milliner born in Galway, Ireland. His mother and father had arrived in Australia in 1848 and 1860, respectively, and married in Melbourne, Victoria, on 11 June 1866, moving to Sydney five years later.[1] Although Lang's father had been born Presbyterian, he later became a Catholic like his wife, and the family "fitted into the normal low social stratum of the great majority of Sydney's Catholics".[2]
The family lived in the inner-city slums for the majority of Lang's early childhood, including for a period on Wexford Street in Surry Hills, where he attended a local school, St Francis Marist Brothers' on Castlereagh Street. His father suffered from rheumatic fever for most of Lang's childhood, and he supplemented his family's income by selling newspapers in the city on mornings and afternoons.[1] In the mid-1880s, due to his parents' poverty, he was sent to live with his mother's sister on a small rural property near Bairnsdale, in the Gippsland region of Victoria, attending for about four years the local Catholic school. Lang returned to New South Wales in the early 1890s to seek employment, aged 14. His first jobs were in the rural areas to the south-west of Sydney: on a poultry farm at Smithfield, and then as the driver of a horse-drawn omnibus in and around Merrylands and Guildford.[2]
Aged 16, he returned to the inner city, working first in a bookstore, and then as an office boy for an accountant. Nairn (1986) writes that Lang's experience in the Sydney slums brought "an intimate knowledge […] of the protean denizens who found shelter there", inculcating in Lang some "real sympathy for them, but above all a determination to avoid their kind of existence, reinforced by a revulsion against the hardships of his own life in a large, generally poverty-stricken family."[3]
During the banking crash of the 1890s which devastated Australia, Lang became interested in politics, frequenting radical bookshops and helping with newspapers and publications of the infant Labor Party, which contested its first election in New South Wales in 1891. At the age of 19 he married Hilda Amelia Bredt (1878–1964), the 17-year-old daughter of prominent feminist and socialist Bertha Bredt, and the step-daughter of W. H. McNamara, who owned a bookshop in Castlereagh Street. Hilda's sister, also named Bertha, was married to the author and poet Henry Lawson.
Lang became a junior office assistant for an accounting practice, where his shrewdness and intelligence saw his career advance. Around 1900 he became the manager of a real estate firm in the then semi-rural suburb of Auburn. He was so successful that he soon set up his own real estate business in an area much in demand by working-class families looking to escape the squalor and overcrowding of the inner-city slums.
Lang continued his political pursuits, soon becoming an Alderman on Auburn Municipal Council and eventually mayor. He was elected as a member of the New South Wales Legislative Assembly in 1913 for the district of Granville, serving as a backbencher in the Labor Party government led by William Holman. When Prime Minister Billy Hughes twice tried to introduce conscription to the country in WWI, Lang sided with the anti-conscriptionist wing of the ALP. The mass defection from the ALP of parliamentarians and supporters who supported the military measure opened up opportunities and Lang positioned himself for advancement. His financial skills led him to become Treasurer in Premier John Storey's Labor Government from 1920 to 1922. Due to the post-World War I financial recession, the state's accounts were in deficit; Lang managed to cut this deficit significantly. From 1920 to 1927, he was a member for the multi-member seat of Parramatta.
After the Australian Labor Party (ALP) lost government in 1922, Lang was elected as Opposition Leader in 1923 by his fellow Labor Party MPs. He led the ALP to victory in the 1925 NSW general election and became Premier.
During his first term as Premier, Lang carried out many social programmes, including state pensions for widowed mothers with dependent children under fourteen, a universal and mandatory system of workers' compensation for death, illness and injury incurred on the job, funded by premiums levied on employers, the abolition of student fees in state-run high schools and improvements to various welfare schemes such as child endowment (which Lang's government had introduced). Various laws were introduced providing for improvements in the accommodation of rural workers, changes in the industrial arbitration system, and a 44-hour workweek. Extensions were made to the applicability of the Fair Rents Act whilst compulsory marketing along the lines of what existed in Queensland was introduced. Adult franchise for local government elections was also introduced, together with Legislation to safeguard native flora and to penalize ships for discharging oil.[citation needed] His government also carried out road improvements, including paving much of the Hume Highway and the Great Western Highway.
Lang established universal suffrage in local governmentelections – previously only those who owned real estate in a city, municipality or shire could vote in that area's local council elections. His government also passed legislation to allow women to sit in the upper house of the New South Wales Parliament in 1926. This was the first government to do so in the British Empire and three years before the 'Persons Case' decision of the Privy Council in London would grant the same privilege to women throughout the Empire. However, his attempts to abolish the appointive upper house of the NSW Parliament, the Legislative Council, were unsuccessful.
After Labor's defeat at the 1927 election, Lang was Opposition Leader again from 1927 to October 1930. He was a member for Auburn from 1927 to 1946. In this period the Great Depression had begun in earnest with devastating effects on the welfare and security of Australia.
In 1930, more than one in five adult males in New South Wales was without a job. Australian governments responded to the Depression with measures that, Lang claimed, made circumstances even worse - cuts to government spending, civil service salaries and public works cancellations. Lang vigorously opposed these measures and was elected in a landslide in October 1930.
As Premier, Lang refused to cut government salaries and spending, a stand which was popular with his constituents, but which made the state's fiscal position more parlous, though the economic state of the six other various Australian governments fared little better during this same period. In the wake of the Great Depression, measures were taken to ease the hardships of evicted tenants together with the hardships facing householders and other debtors battling to meet repayments.[citation needed] He passed laws restricting the rights of landlords to evict defaulting tenants, and insisted on paying the legal minimum wage to all workers on relief projects.
At an economic crisis conference in Canberra in 1931, Jack Lang announced his own programme for economic recovery. The "Lang Plan" advocated the repudiation of interest payments to overseas creditors until domestic conditions improved, the abolition of the Gold standard to be replaced by a "Goods Standard" where the amount of money in circulation was linked to the amount of goods produced, and the immediate injection of £18 million of new money into the economy in the form of Commonwealth Bank of Australia credit. The Prime Minister and all other state Premiers rejected the plan.
Lang was a powerful orator, and during the crisis of the Depression he addressed huge crowds in Sydney and other centres, promoting his populist program and denouncing his opponents and the wealthy in extravagant terms. His followers promoted the slogans "Lang is Right" and "Lang is Greater than Lenin." Lang was not a revolutionary or even a socialist, and he loathed the Communist Party, which in turn denounced him as a social fascist.
On 19 March 1932, Lang opened the Sydney Harbour Bridge. Lang caused some controversy when he insisted on officially opening the bridge himself, rather than allowing the Governor, the King's representative in NSW, to do so. He delivered what has come to be regarded as a landmark speech in Australian political history during the Opening, citing the theme that the completion of the Sydney Harbour Bridge was analogous to the history, development and dreams of the Australian nation and its people. It may be inferred that this speech depicted Lang's personal vision of the past, present and future of New South Wales and Australia's place in the British Empire and world, (to read this speech, refer to 'Stirring Australian Speeches', edited by Michael Cathcart and Kate Darian-Smith). Just as Lang was about to cut the ribbon to open the Sydney Harbour Bridge, Captain Francis de Groot, a member of the paramilitary New Guard movement, rode up and broke the ribbon. The New Guard also planned to kidnap Lang, and plotted a coup against him during the crisis that brought Lang's premiership to an end.
Early in 1931, Jack Lang released his own plan to combat the Depression; this became known as "the Lang Plan". This was in contrast to the "Melbourne Agreement" which all other State Governments and the Federal Government had agreed to in 1930. Key points of the Lang Plan included the reduction of interest owed by Australian Governments on debts within Australia to 3%, the cancellation of interest payments to overseas bondholders and financiers on government borrowings, the injection of more funds into the nation's money supply as central bank credit for the revitalisation of industry and commerce, and the abolition of the Gold standard, to be replaced by a "Goods Standard," whereby the amount of currency in circulation would be fixed to the amount of goods produced within the Australian economy. The banks had indicated that if he paid the interest they would advance him an additional amount which was greater than the interest, thus giving him a positive cash flow.
Lang opposed the Premiers' Plan agreed to by the federal Labor government of James Scullin and the other state Premiers, who called for even more stringent cuts to government spending to balance the budget. In October 1931 Lang's followers in the federal House of Representatives crossed the floor to vote with the conservative United Australia Party and bring down the Scullin government. This action split the NSW Labor Party in two – Lang's followers became known as Lang Labor, while Scullin's supporters, led by Chifley, became known in NSW as Federal Labor. Most of the party's branches and affiliated trade unions supported Lang.
Since the Commonwealth Government had become responsible for state debts in 1928 under an amendment to the Constitution, the new UAP government of Joseph Lyons paid the interest to the overseas bondholders, and then set about extracting the money from NSW by passing the Financial Agreement Enforcement Act 1932, which the High Court held to be valid. Lang then contended that the Act was rendered null and void by contravening the 1833 prohibition of slavery throughout the British Empire; the Premier held that the actions of the Lyons government deprived the State of New South Wales of paying the wages of State employees and that this necessarily constituted an (illegal) state of slavery.
In response, Lang withdrew all the state's funds from government bank accounts and held them at Trades Hall in cash, so the federal government could not gain access to the money. The Governor, Sir Philip Game, a retired Royal Air Force officer, advised Lang that in his view this action was illegal, and that if Lang did not reverse it he would dismiss the government. Lang stood firm, and on 13 May 1932 the Governor withdrew Lang's commission and appointed the UAP leader, Bertram Stevens, as premier. Stevens immediately called an election, at which Labor was heavily defeated.
Gerald Stone, in his book 1932, states that there is evidence that Lang considered arresting the Governor to prevent the Governor from dismissing him, (which Lang admitted in his own book, The Turbulent Years). The possibility was sufficiently high that the armed forces of the Commonwealth were put on alert. Michael Cathcart and Andrew Moore, among others, have put forward the possibility that such a clash would have seen the Commonwealth Armed Forces fighting the New South Wales Police.
This was the first case of an Australian government with the confidence of the lower house of Parliament being dismissed by a Vice-Regal representative, the second case being when Governor-General Sir John KerrdismissedGough Whitlam's government on 11 November 1975. Game himself felt his decision was the right one, despite the fact that he had no personal animosity towards Lang. On 2 July 1932 Game wrote to his mother-in-law: "Still with all his faults of omission and commission I had and still have a personal liking for Lang and a great deal of sympathy for his ideals and I did not at all relish being forced to dismiss him. But I felt faced with the alternative of doing so or reducing the job of Governor all over the Empire to a farce."[4] Lang himself, despite objecting to his dismissal conceded afterwards that he too liked Game, regarding him as fair and polite, and having had good relations with him.[5]
There is an existing disagreement since this time whether the Governor, no matter how personally convinced he'd been one way or the other, had actually had authority to dismiss Lang on the grounds which he had. It is contended that the ability to determine criminal guilt, before and under the law, belonged solely to the established justice system, not to the Governor. Lending weight to this view is the stipulation of clause 39 of the 1215 Magna Carta (clause 29 in the 1297 version), which had established this very precept in the United Kingdom. Further, the 'Letters Patent' of the New South Wales Governor explicitly stated that the Reserve Powers of the Governor could only be expressed in any way which the Monarch of the United Kingdom could express the same.
Lang continued to lead the Labor Opposition, although the NSW Branch of the ALP remained separate from the rest of the party. The UAP won the elections of 1935 and 1938. After this third defeat, the Federal Labor forces began to gain ground in NSW, as many union officials became convinced that Labor would never win again in the state while Lang remained leader. Lang was ousted as NSW Opposition Leader in 1939 and was replaced by William McKell, who became Premier in 1941.
Lang was expelled from the ALP in 1942, and started his own parallel Labor Party, called the ALP (Non-Communist), but this time with only minority support in the NSW party and unions. Through the 1940s, he railed against the dangers of communism as a 'Cold War warrior'. He remained a member of the Legislative Assembly until 1946, resigning to stand for the Division of Reid in the Australian House of Representatives. His state seat of Auburn was won by his son James Lang at a by-election. Jack Lang's victory in Reid was unexpected; he was elected on a minority of the votes thanks to preferences given to him by the Liberal Party. In federal parliament, he is often cited as being the most effective of the opposition to the government of his old rival, Prime Minister Ben Chifley[citation needed], despite voting for the latter's Bank Act in 1947. In 1949 he was defeated and never held office again, despite a bid to be elected to the Senate in 1951.
Lang spent his long retirement editing his newspaper The Century, and wrote several books about his political life, including The Great Bust, I Remember and The Turbulent Years. He grew increasingly conservative as he grew older, supporting the White Australia Policy after the rest of the labour movement had abandoned it. In I Remember he wrote: "White Australia must not be regarded as a mere political shibboleth. It was Australia's Magna Carta. Without that policy, this country would have been lost long ere this. It would have been engulfed in an Asian tidal wave." To the end of his life, he proudly proclaimed that "Lang was Right." Lang also spent time visiting Sydney schools recounting recollections of his time in office to his young audience. Lang gave a number of lectures at Sydney University circa 1972–1973, at which he discussed his time in office and other topics such as economic reform. His address given on 1 July 1969 to the students of Sefton High School is available on tape at the Mitchell Library. He was re-admitted to the Labor Party in 1971, aided by his young protege Paul Keating.
Lang died in Auburn in September 1975, aged 98, and was commemorated with a packed house and overflowing crowds outside Sydney's St. Mary's Cathedral at his Requiem Mass and memorial service. His funeral was attended by prominent Labor leaders including then Prime Minister Gough Whitlam. He was buried at Rookwood Cemetery, Sydney.[6]
Freudenberg, Graham (1991). Cause for power: the official history of the New South Wales Branch of the Australian Labor Party (paperback). Sydney: Pluto Press in association with the Australian Labor Party. New South Wales Branch. p. 297. ISBN0-949138-60-6.
Moore, Andrew (1989). The secret army and the Premier: conservative paramilitary organisations in New South Wales 1930-32 (paperback). Kensington, NSW: New South Wales University Press. p. 312. ISBN0-86840-283-4.
Robertson, John (1974). J.H. Scullin: A Political Biography. Nedlands, W.A.: University of Western Australia Press. p. 495. ISBN0-85564-074-X.
Robinson, Geoffrey (1992). How Labor governed: social structures and the formation of public policy during the New South Wales Lang government of November 1930 to May 1932 (Thesis). Melbourne: Monash University, Dept. of History.
Henderson, Anne (September 2007). "All quiet on the civil war front: it was the civil war that never happened. Or was it? [Where is the evidence to support the theory that Australia was in a pre-civil war condition in 1931 and 1932?]". The Sydney Institute Quarterly (The Sydney Institute) (31): 6–12. ISSN1441-4074.
Walker, Robin (November 1986). "Mr. Lang's dole: the administration of food relief in New South Wales, 1930-32". Labour History (Australian Society for the Study of Labour History) (51): 70–82. ISSN0023-6942.
|
{
"pile_set_name": "Pile-CC"
}
|
holder
PARIS: Nine-time champion Rafael Nadal and title-holder Novak Djokovic surged into the French Open quarter-finals for a record-equalling 11th time on Sunday (Jun 4), but reigning women's champion Garbine Muguruza crashed out in the last 16.
Bui Kim Hung is a student at the National University of Civil Engineering but he may have to drop out to take care of his young brother after his mentally-ill father killed both his mother and grandmother.
|
{
"pile_set_name": "Pile-CC"
}
|
Search
SCUBAPROSCUBACLEAR ANTIFOG DROPS
SKU24090000
The easy bacteria free solution to mask de-fogging. The defogging fluid puts a liquid film on the surface of the mask lens to help prevent the fogging so you can see clearly while underwater. Simply drop the scuba clear solution into your mask before every dive, rub in and rinse for crystal clear vision throughout the entire dive.
|
{
"pile_set_name": "Pile-CC"
}
|
Q:
Android: How to keep App Locale work independently from System Locale?
I have an issues related to display language. I am able to change the language inside the application ("en" for English and "ja" for Japanese) independently from OS system.
However, the problem is that when the application is in "ja" if users change the System language manually (not "en" or "ja") than my application auto-change the language to default ("en"). I want to make locale of my application stand alone, whatever language users change manually, the language of application still remain the same as when they log out.
EDIT
There are some useful links but they still cannot solve my problem. For example:
Change language programatically in Android
Could you give my any suggestion to do it?
Thank you in advance!
A:
Try this one:
import java.util.Locale;
import android.os.Bundle;
import android.app.Activity;
import android.content.Intent;
import android.content.res.Configuration;
import android.content.res.Resources;
import android.util.DisplayMetrics;
public void setLocale(String lang) {
myLocale = new Locale(lang);
Resources res = getResources();
DisplayMetrics dm = res.getDisplayMetrics();
Configuration conf = res.getConfiguration();
conf.locale = myLocale;
res.updateConfiguration(conf, dm);
Intent refresh = new Intent(this, AndroidLocalize.class);
startActivity(refresh);
finish();
}
ADDED:
@Override
public void onConfigurationChanged(Configuration newConfig) {
newConfig.setLocale(yourLocale);
super.onConfigurationChanged(newConfig);
}
ADDED(2):
You have to set android:configChanges="layoutDirection|locale" in order to trigger onConfigurationChanged() when you change Locale.
I cannot fully understand why this happens, maybe there are some RTL languages...
|
{
"pile_set_name": "StackExchange"
}
|
Q:
Why is BufferedWriter not writing to file?
I have this code:
String[] strings = {"Hi" , "You", "He", "They", "Tetrabenzene", "Caaorine", "Calorine"};
File file = new File("G:\\words.txt");
FileWriter fWriter;
BufferedWriter bWriter;
try {
if((file.exists())) {
file.createNewFile();
}
fWriter = new FileWriter(file.getAbsoluteFile(), true);
bWriter = new BufferedWriter(fWriter);
//Convert Result objects to JSON and write to file
for(int j = 0; j < strings.length; ++j) {
bWriter.write(strings[j]);
bWriter.newLine();
System.out.println("done");
}
}
catch(IOException e) {e.printStackTrace();}
I have pretty much the same code 2 or 3 times before this and BufferedWriter is writing perfectly. But for some reason when I get to this code it doesn't write. I've been looking for things that might be wrong but I can't change something and test it quickly as the program takes 10 minutes just to get to this part.
Also, the program prints "done" to the console so I know it is going into the for loop.
Any ideas as to what I'm doing wrong?
A:
Call bWriter.flush() when you want your data to actually be flushed to your file on disk.
Or just call bWriter.close() when you're done working with your writer.
The bWriter.close() call will call bWriter.flush() internally.
|
{
"pile_set_name": "StackExchange"
}
|
If you are new to the forums, you must register a free account before you can post. The forums have a separate registration from the rest of www.chronofhorse.com, so your log in information for one will not automatically work for the other. Disclaimer: The opinions expressed here are the views of the individual and do not necessarily reflect the views and opinions of The Chronicle of the Horse.
I don't think there are pros who still get physically sick/nervous. I just don't. Can you image how long you would last at your job if that's how you felt going into work every day?
I am an attorney and sometimes feel this way when I have to go into court (depending on what the matter is). It has improved the longer I have been in practice, but I don't think it will ever truly go away. No one can tell, though, and it does not negatively impact my performance in court. When court is called, I'm suddenly fine. Maybe it is similar for the pros?
YES! George Morris struggled with his nerves for a long time. And, while I am no George Morris, I used to be a very timid rider. It actually proved useful to me in the show ring...guess I ride well under pressure.
"A horse gallops with his lungs, perseveres with his heart, and wins with his character." - Tesio
I never had the degree of anxiety that you're dealing with, but feel like I overcame my own nerves by really owning them. Instead of trying not to be afraid or ignore the feeling, I put it under a microscope. What exactly was I worrying about? How could I keep that one thing from ruining the rest of my ride?
Outside of riding, when you get tense, how do you resolve it? Does following a consistent routine help? Do you need to break down the stressful situation into smaller chunks so you can manage better? Does it help you to move on if you take a moment to freak out first or does someone to telling you its ok help?
Once you've got a sense of how you deal best with other situations, then apply that to your riding. What triggers the tension? Is it the course of jumps you're asked to do or feeling rushed through the flatwork? Would breaking down exercises help? What if you knew your horse had schooled through the exercise earlier in the day?
When you've really thought this through, set up a time to talk to your trainer when they can focus on the conversation. Let them know what stresses you out and how you handle it outside of riding. Agree on a plan for when you get nervous in a lesson so you can quickly communicate to them and they can help you to work through it. It could be anything from trotting the "scary" jump in a course as a single fence before jumping the full course to spending more time on a flat exercise before jumping. Also talk to them, if they could change one thing about the way you ride when you're nervous, what would it be. Then when the anxiety starts, focus on that one thing.
After your lesson, talk (and think) about the successes and what you accomplished. If you were nervous and overcame it, talk about that but avoid the "did you see what so and so did, I'd never be brave enough" conversations. Put all your mental energy into the progress you're making so you can really own it, instead of focusing on the anxiety inducing what ifs.
It took a while, like a year or two, but now I can use that gut feeling to figure out, is there a real safety issue that my body is warning me about (like pushing too hard when overtired) or is it unjustified and to feel comfortable deciding that sometimes quitting while I'm ahead isn't running scared it's making sure that I really enjoy my time at the barn.
I also ditto the thought of narrowing the focus. When I'm at a show on course, instead of allowing myself to freak out over EVERYTHING )are my heels down? Am I balanced? Am I straight? Oh my god, I'm going to chip! Crap, the lead!), I will pick ONE thing at a time in my head and put my focus into it. So It goes like this:
Circle - pick up lead. Lead lead lead.
Head to first jump- straight straight straight (I find when I'm straight to the first jump, everything else falls better into place).
Heading to second jump - feel the stride, feel the stride, feel the stride.
Turning to next line- lead change lead change, lead change.
Corner corner corner corner. USE THE CORNER
Feel the stride, feel the stride, feel the stride, feel the stride JUMP
etc. I find when I worry about the "bigger" things, the other things fall into place. I KNOW what position I need to be in and that I need to have good equitation, and I find that if I tune out the nagging thoughts about it, everything tends to come out OK. Now, it may not be perfect, but it will never be the disaster I think it's going to be in my mind.
Focus on ONE thing at a time in your head, obsessively, and it tends to take some nerves away.
Try some time with a GOOD sport psychologist (if just in riding), or a psychologist if it is across the board. You have to have a plan about how to replace your behaviors. A person with no fear is a fool or a liar, the question is what you do with your thoughts. Channel the AWARENESS into learning. Learning a better seat (riding w/o stirrups/planning a line/etc). Your mind is either OCCUPIED with what to do messages, or introverted to #$*($& which is merely a distraction. So, choose the right thoughts, use the focus which comes with the simulation, it is a positive (channel your inner GM!). And it leads me to wonder how PROGRESSIVELY you have been presented with equitation/caveletti/jumping (and jumping bigger is just a simple progression). (Interestingly enough, 2'6" used to be a hack class, and 3' was only for green horses, not riders. Jumping smaller fences is actually more difficult for good equitation.) Learning to ride is one thing, learning to train is another. IF you work progressively both the horse and the rider will have NO problems.
Are there any good books out there on sports psychology? I got dumped pretty good last April (trip to the ER with CT scan and bad concussion), and although I've been riding since then, something has changed in my mentality and I haven't been quite able to break through it. I'd love to attend a sports psychology clinic too...ideas?
I had really bad nerves after a bad fall at a show. It didn't make any sense at all, because I have the most honest and wonderful horse on the planet. Took me two shows but I worked out of it and love showing again. I get so energized and excited and ready to go. Maybe I've just finally got all my butterflies flying in the same formation, but I don't feel bad jitters... just jitters! I could never imagine wanting to go pro or "make it" if I always had nerves like I did for the two shows after my bad fall, and from the sound of it, what I had was much more mild than what many people experience.
[QUOTE=GingerJumper;6775016]I had really bad nerves after a bad fall at a show. It didn't make any sense at all, because I have the most honest and wonderful horse on the planet. Took me two shows but I worked out of it and love showing again. I get so energized and excited and ready to go. Maybe I've just finally got all my butterflies flying in the same formation, but I don't feel bad jitters... just jitters! I could never imagine wanting to go pro or "make it" if I always had nerves like I did for the two shows after my bad fall, and from the sound of it, what I had was much more mild than what many people experience.[/QUOTE
I get terribly nervous, have spent many sleepless nights before shows! The things that help me.... Not trying to pretend I'm not afraid - I tell my trainer how I feel and we can talk about it, make a joke about it, etc. Trusting my horse and thinking about her good points, realizing SHE is only afraid of things when I tell her to be (she's so much braver than I am!) and trusting her to try for me. Making myself go, go, go - to schooling days, clinics, lessons at at another farm (trainer and I go together to ride with a BNR) - so I feel like we've done our homework, horse has the background to do the job, etc. I get so nervous before ANY trip that I wonder why I do it to myself - but then the feeling when it's over and we've done well is more than worth it.
Another great thing my trainer has done is breaking the whole thing into bite size pieces. We get to the show, set up, horses go in stalls, we go take a walk to check in and see the rings/fences, and see that the puddle jumpers are not doing anything Olympic (in my imagination, there are liverpools and triple bars in the 2'6" division)... If everything is good, we make our schooling/warm up plan and go tack up. If everything is good, we start jumping a few warm up fences. If everything is still good, we canter/jump a few lines. If everything is STILL good, we synchronize watches to get ready for an actual class. It's always, "Just trot around. You don't have to do anything but trot around." "Okay, now just jump this warm-up fence. That's all you have to do." I can handle my nerves for one thing at a time. I have a hard time handling my nerves for an entire weekend of showing, but I can do one thing at a time. (When we are showing frequently, sometimes I can handle two or three things at a time! It's a miracle!)
I used to teach lessons and ride a few horses for other people. Weirdly, my nerves are easier to deal with when it's someone else's horse? I don't understand why. But overall I was too nervous to keep it up. I loved teaching and my riders did well, but I just about died of anxiety every time one of them trotted into the ring. I can't imagine a successful pro having that kind of nerves - it makes the job no fun.
I tell myself that we all have our strengths - a lot of people are afraid of public speaking, and I love it - doesn't make me nervous for a second! I wish I could trade in my public speaking bravery for horse show bravery, though.
Focus on ONE thing at a time in your head, obsessively, and it tends to take some nerves away.
A "spin-off" of this that I always think about: make everything take AS LONG as possible. Whenever you run through a lead change or gun it at a jump, your mind is taking over and running faster than the actual events that are happening! I've found that if I think "I am going to make this lead change last FOREVER" or "There are a MILLION strides between me and that single oxer," it slows everything down in my head.
A "spin-off" of this that I always think about: make everything take AS LONG as possible. Whenever you run through a lead change or gun it at a jump, your mind is taking over and running faster than the actual events that are happening! I've found that if I think "I am going to make this lead change last FOREVER" or "There are a MILLION strides between me and that single oxer," it slows everything down in my head.
YES! When I am so nervous that I am in a hurry to get things over with, that's when I royally screw up. Trying to make it last a long time is very good advice.
Read the George Morris article that was posted not too long ago. He talked a bit about his nerves and how they translated into better riding for him. Perhaps you could write a letter to him asking about he deals/dealt with his serious nerves.
As a junior, I had the "bad kind" of nerves—where you go in and sort of freeze and go on autopilot. Now, for some reason, as an amateur, my nerves have transformed themselves into that "edge" I was talking about—I'm still a little nervous at home, but the second I step into the show ring, that all goes away and I'm able to make calculated decisions and process and really ride. I wish I could explain what clicked in my head, but I can say that showing and riding A LOT makes a big difference.
I am kind of the same way. As a pony kid, I would get so nervous, I just wouldn't think. I had a lot of off courses to my record. I didn't show for over 20 years, but while I wasn't showing, I spent a lot of time grooming for people showing at levels much higher than I ever had, so I had a lot of time to watch them and observe them ride into the ring like it was no big deal. Then I started to think it was no big deal. Preparing for a horse show now is so old hat that it doesn't feel like a big deal to me, which keeps my anxiety at bay because I am not worrying about my tack being scrupulously clean or if I remembered a pair of spurs.
In addition, showing youngsters where there are really no expectations has also kept my expectations/anxiety at bay. I set a goal (let's say straightness) and I go in and try to achieve that.
Originally Posted by Tha Ridge
A "spin-off" of this that I always think about: make everything take AS LONG as possible. Whenever you run through a lead change or gun it at a jump, your mind is taking over and running faster than the actual events that are happening! I've found that if I think "I am going to make this lead change last FOREVER" or "There are a MILLION strides between me and that single oxer," it slows everything down in my head.
Totally worked for me. When I was younger, everything seemed to come up so "fast", which got me in trouble. Now, when I catch myself speeding up my thoughts or actions, I take a deep breath to slow things down and reset my perspective.
I am kind of the same way. As a pony kid, I would get so nervous, I just wouldn't think. I had a lot of off courses to my record. I didn't show for over 20 years, but while I wasn't showing, I spent a lot of time grooming for people showing at levels much higher than I ever had, so I had a lot of time to watch them and observe them ride into the ring like it was no big deal. Then I started to think it was no big deal. Preparing for a horse show now is so old hat that it doesn't feel like a big deal to me, which keeps my anxiety at bay because I am not worrying about my tack being scrupulously clean or if I remembered a pair of spurs.
Hm, we are actually very similar then—after showing a lot as a junior and my first few years of college, I went out and worked for a BNT on the West coast. I showed some, but mostly prepped horses for the big eq kids. A lot of riding time at shows, but not actually showing, and a lot of time spent on the sidelines. Ever since my return to showing following all of that, I've been tremendously calm and collected at my own shows. There must be something to spending time out of the saddle, observing, too!
This might sound a little backward, but what helped me with my show nerves was being unprepared. The last show I was in (so this might not actually permanently fix you-haven't had the opportunity to show since then) was when I was in college. I had been home for the summer for 2 weeks and my friend was going to a local show and told me I should load my horse up and come along. We were both out of shape and hadn't had lessons since Christmas. I knew we were totally unprepared so I stopped caring and relaxed. It was the most fun I've ever had at a show and we ended up with reserve champion for our division, which is my highest show accomplishment (so I'm obviously no BNR!).
Another thing that helped me when I was taking regular lessons most recently was getting a trainer to tell me to do it. She's a no-nonsense trainer, and I'm terrible at telling people no. Turns out it was a good combination! She would set a course for me on XC and I would look around and think there was no way, and that the things she was pointing to were giant (definitely not giant, but to me they were), but I didn't want to tell her no or that I was afraid so I'd just suck it up and do it. It worked wonders and despite my fears, we didn't die I knew she wouldn't tell me to jump something I wasn't ready for. I would have told her no/refused if I had felt otherwise, but I trusted her and put on my big girl panties and those were some of my most fun and exciting rides!
So for me it's all about trusting your horse, your trainer, and realizing it's all about having fun. We'll see if I remember that someday when I start showing again
Really happy to read all the suggestions... I was having a lot of fear issues with jumping, seemed to get better with it, and then had a couple of falls when my horse started stopping again and I went right back to square one. I fortunately have worked with all good instructors recently who don't overface me, and let me do stuff slowly, so I'm starting to regain a little bit of confidence again. But it's such a primal and unreasoning sort of fear that it's really hard to try to get past it--I can sit on the couch and think about jumping and I'm suddenly having a panic attack.
I had really bad nerves after a bad fall at a show. It didn't make any sense at all, because I have the most honest and wonderful horse on the planet. Took me two shows but I worked out of it and love showing again. I get so energized and excited and ready to go. Maybe I've just finally got all my butterflies flying in the same formation, but I don't feel bad jitters... just jitters! I could never imagine wanting to go pro or "make it" if I always had nerves like I did for the two shows after my bad fall, and from the sound of it, what I had was much more mild than what many people experience.[/QUOTE
If you weren't scared, you wouldn't be human
You mean if I wasn't scared after a bad fall, or are you reffering to the rest of it?
I used to get really nervous at the in-gate. Butterflies, the whole nine yards. As show season would go on, I'd be less nervous (I guess I was getting used to it, and I was getting good results....I think a lot of nervousness stems from "Oh man, I'm doing this in front of a ton of people what if I screw up? What if I'm not as good as I think I am? Took me screwing up a few times and eating a few jumps to realize that everyone does it and no one really makes that big a deal out of it! A few blue ribbons didn't hurt, either.)
Now I have found that my main source of nervousness is the fact that I'm a re-rider and it's 10 years later, and I'm SUPER out of shape. I'm all brave and think to myself "EFF IT! I'm gonna go ride the heck out of this horse and I MAY EVEN JUMP A FEW JUMPS" until I actually start riding and then my confidence goes out the window because it's there in black and white how out of shape I am and how quickly I'm winded. I'm working on it, though.
If only I could figure out how to not have feelings when I'm on top of a horse, just like how I don't have feelings in the rest of my life.....
|
{
"pile_set_name": "Pile-CC"
}
|
DESCRIPTION:
After having a bad start to the day with the taxi breaking down, I was off to do some private car hire in my car, the same thing really just not in my taxi, but it's no good for shagging. After driving around my usual stops I noticed a customer outside one of the clubs in the area so pulled in to see if she needed a taxi service. As I suspected she did, she was a wife to a very rich local man but she was not just after a shopping trip she wanted cock and lots of it. She offered me money to give her a good fuck as her husband can't get hard. Not only was she offering me money to fuck her but she was also very sexy and I could not wait to show this posh chick a good fucking.
... Show more
Show less
|
{
"pile_set_name": "OpenWebText2"
}
|
Its my favourite time of year at the cinema, awards season. I decided to go and see this movie blind with my only knowledge being that it was chosen as one of the official selections at Sundance.
The film follows Lee Chandler (Casey Affleck) a despondent, irritable loner who works as a handyman for a Boston apartment block. One damp winter day he gets a call summoning him to his hometown. His brother has died and he’s been named guardian to his 16-year-old nephew. As if losing his only sibling and doubts about raising a teenager weren’t enough, he is forced to unwillingly revisit his troubled past.
First of all I’m going to say that movies like these remind me of why I love cinema so much.
I know that in the media Casey Affleck has been getting a large amount of praise about his performance described frequently as a tour-de-force (and it is) but what makes his portrayal of Chandler so great is the realism he brings to his character. The overall emotive effect of the movie is nuanced by how he completely nails his performance and the subtlety of Affleck’s character development throughout is something that most, especially introverts will find relatable to their own real life experience.
I don’t want to give anything away about what actually happens within the story; the writing is so tight, most of the storytelling devices are pulled off to perfection and mainly the way that the many different relationships throughout are portrayed is greatly true to life.
There are no jaw-dropping, blockbuster moments in Manchester By The Sea but I really feel that this movie is one that everyone should watch at least once, as a solemn reminder that we are all people with our own highs, lows and stories to tell.
A reminder that you are not alone.
Platform I Watched On: Cinema Release
For fans of: Room (2015), Drama, Oscar worthy performances.
My rating: 92 popcorns
IMDB: 8.5
Rotten Tomatoes: 97%
|
{
"pile_set_name": "OpenWebText2"
}
|
Relationship between two serotypes of avian mycoplasmas by serological and DNA-DNA hybridization tests.
Genetic relatedness among eight strains of avian mycoplasmas representing two serotypes, C-O and D-P, was determined using DNA-DNA hybridization tests. One strain from each serotype was found to be misclassified. The average percent homology between the two serotypes was 34.1% at 56 C and 18.6% at 66 C. The average homology among the four C-O organisms was 99.6% at 56 C and 93.5% at 66 C, while among the four D-P strains the average homology was 91.6% at 56 C and 92.2% at 66 C. These relationships correlated with antigenic cross-reactions observed in complement fixation (CF) and fluorescent antibody (FA) tests. Thermal denaturation of the pure DNA from each strain resulted in base compositions of 25.0 to 28.2 moles percent guanosine plus cytosine (GC). Thermal denaturation of selected hybrid DNA revealed a linear relationship between percent homology and delta melting temperature (1 C delta Tm for 15% difference in percent homology). Thirty-four percent homology between the organisms of C-O and D-P serotypes indicates a minor antigenic relationship; thus, it is recommended that these two serotypes be retained as separate groups.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Q:
Metaphor for the concept of shell?
I'm finding myself helping out some classmates in my computer science class, because I have prior development experience, and I'm having a hard time explaining certain things like the shell. What's a good metaphor for the shell in the context of the Terminal on Mac, contrasted with a remote shell via SSH?
A:
Hey you speak with the computer!
(Quote from a child who discovers the shell.)
“The shell” is basically a language the computer can recognize, obey to, and reply to if asked.
Local or not makes no difference. Think of a remote shell as an equivalent of communication over the phone.
A:
A shell is the interface through which a user interacts with a computer. The Windows GUI itself is a shell. A command line shell requires the user to type commands as opposed to point 'n click.
A:
Put simply, a terminal is an I/O environment for programs to operate in, and a shell is a command processor that allows for the input of commands to cause actions (usually both interactively and non-interactively (scripted)). The shell is run within the terminal as a program.
There is little difference between a local and remote shell, other than that they are local and remote (and a remote shell generally is connected to a pty, although local shells can be too).
|
{
"pile_set_name": "StackExchange"
}
|
/// Copyright (c) 2012 Ecma International. All rights reserved.
/**
* @path ch11/11.13/11.13.2/11.13.2-6-7-s.js
* @description Strict Mode - SyntaxError is thrown if the identifier eval appear as the LeftHandSideExpression of a Compound Assignment operator(>>=)
* @onlyStrict
*/
function testcase() {
"use strict";
var blah = eval;
try {
eval("eval >>= 20;");
return false;
} catch (e) {
return e instanceof SyntaxError && blah === eval;
}
}
runTestCase(testcase);
|
{
"pile_set_name": "Github"
}
|
He is 42, and she is half his age.
Carson*, a writer, met his sugar baby, Amy*, an accountancy undergraduate at a Malaysian university, through TheSugarBook, a controversial dating platform that connects women with men who are willing to pay for their upkeep in exchange for companionship.
Despite their age difference, and the fact that one party entered the relationship solely for financial gain, Carson and Amy have become fast friends.
In fact, Amy’s frankness from the start – she told Carson that she was interested only in his bank account – is one of the reasons why Carson agreed to the relationship.
While some might have their reservations about sugar dating, for Carson it is the only way he can meet women. The Malaysian was involved in a car accident 14 years ago, which resulted in him becoming paralysed from waist down.
“For me to date regularly is very difficult. (The chance of me) meeting girls the conventional way is almost zero and some of my friends have tried to set me up before but because of my disability, nothing came to fruition,” he told Yahoo Lifestyle Singapore via e-mail.
While TheSugarBook has been criticised – even in Parliament – for potentially exposing women to emotional and sexual exploitation, it is not the only sugar dating site available here.
A quick search online turned up similar sites such as Seeking Arrangement, aSugarDating and SugarDaddyMeet. Users also post advertisements on classifieds platforms such as Locanto for the purpose of sugar dating.
Since becoming a sugar daddy a decade ago, Carson has had eight “sugar babies”, whom he characterises as women he sponsors and has had close relationships with.
He first got into sugar dating through a chance encounter with a fellow sugar daddy.
He recounted that he was in Germany undergoing cell therapy in his early 30s when he met a Thai businessman in his 40s, who was also a patient. When the two became friends, Carson observed that a “gorgeous and young Ukrainian girlfriend” would visit the businessman at his villa. He also observed that the pair had “camaraderie and rapport”.
Once, the pair invited Carson to join them for coffee and the Ukrainian woman ended up introducing a female friend to Carson.
“It may come off sleazy but what happened was really natural. A couple with a significant age difference hangs out with their single male friend and they just wanted to set me up with a single female friend.
“Just so happens that the female is slightly financially strapped and could use some assistance in that part,” he said.
On why he entered into a sugar dating relationship with Amy, his current and only sugar baby, Carson said, “I am always upfront about my condition in every way possible – there’s no way to hide it (that I am in a wheelchair) and I entered into an arrangement with Amy as a convenient way to ease my loneliness and to have a female companion.”
He recounted that when he revealed his condition to Amy on TheSugarBook during the first time they talked, she asked why Carson’s physical handicap would affect her as she was more interested in his “bank balance”.
“I loved it. So straightforward and sans pity… What can I say, I was hooked,” Carson said, adding that he found her response “very cute”.
He added that “she didn’t flinch a single bit when we first met up for coffee” and “she’s always been kind and friendly to me”.
For Amy, she has had a total of five sugar dating relationships with men who are significantly older than her – including Carson and another sugar daddy, Matthew, whom she is also currently seeing.
“When I was 19, I dated someone who was 32 and I met him through cell group. It was a relationship or sugar relationship – it’s up to you to judge – but he did buy me a lot of gifts as he was clearly better off than I was at 19 with no job,” she recalled.
Amy added that she “casually” dated a few men she met through TheSugarBook, but that was before she got involved with Carson and Matthew.
Quid pro quo relationship
While sex seems to be an unspoken given in typical sugar dating relationships, Amy and Carson’s relationship is different as such physical intimacy is out of the equation.
It is still transactional, however, as Carson gives Amy pocket money of RM2,000 (S$672) a month. This money goes toward her university expenses, clothing, make-up and the occasional flight home to visit her family, who do not reside in Kuala Lumpur, where she is studying.
Money aside, Carson also enriches Amy’s mind by introducing her to the classic works of authors such as George Orwell and Scott Fitzgerald.
Story continues
|
{
"pile_set_name": "OpenWebText2"
}
|
Add these Regal Favor Collection cross themed frames to an occasion that's nothing short of a blessing. For religious occasions or any event where you want to add a memento that's fits for a king or queen, this elegant frame is a spirited choice. The...
|
{
"pile_set_name": "Pile-CC"
}
|
Archives for January 2011
Yes, mobile is regional. Loved this piece in Moconews about how Newscorp is handling mobile content in India. It demonstrates the need to fully understand the local market when offering mobile content services internationally. While mobile internet growth in North America has been substantially boosted by increased smartphone penetration, the story is not so simple in other markets. 700 Million handsets in India… That’s a sizable potential audience for any content play.
Using the voice channel creates some interesting opportunities for content owners. When we were building a mobile web presence for NPR several years ago (their first), we used an IVR (Interactive Voice Response) system to allow the NPR audience to listen to radio broadcasts on-demand via the phone’s voice channel. At the time, only a small percentage of U.S. mobiles could download or steam audio. Those that could were typically not tied to unlimited data plans. We used IVR as a clever work-around because audio is central to NPR’s brand. This provided simple access to audio on mobile for ANY NPR listener. Further, it required only a small change in user behavior and had no learning curve or technical hoops to jump through.
We’re currently working with a global content owner to help them leverage their content via mobile in international territories. An understanding of the local population, their culture, their habits and prevailing technologies (networks, handsets and billing options) will be critical to their success.
That success, may (and likely should) take on different forms and leverage varying technologies in each territory. Yes, mobile is a global phenomenon, but a one-size fits all approach will leave much of the audience behind.
Apple’s attitude and efforts need to be friendlier to content creators or the creators will find other routes to market or will stop creating content. If the walled garden isn’t fertile, why plant seeds? They certainly don’t need apple’s help to give away their content.
I love this idea as a way to fix some of the fair-use issues that come up with digital media. This seems like a great solution. Why not let a person promote your content by making it easy to share (loan)? I always bought my vinyl (I’m dating myself, I know) but made audio cassettes for friends in order to turn them on to something new.
|
{
"pile_set_name": "Pile-CC"
}
|
I teamed up with my friends at Whole Foods Market NYC to bring you The Crunchy Radish Cinco Series, a healthier take on a festive holiday.
Most of my friends know tequila is my drink of choice. Blanco, rocks, splash of fresh lime, maybe a little club soda...a "skinny girl" rose to fame with something kind of similar. For this Cinco de Mayo Celebration, I made a little variation on my favorite drinking theme. Muddled cucumber and fresh raspberries add a festive and bright touch to zesty lime juice, bubbly club soda, and smooth tequila. Light, refreshing, delicious. The perfect way to get your fiesta started.
Drink responsibly Cinco revelers. Cheers!
Raspberries + Cucumber Tequila Spritzer
Serves 1
3-4 raspberries, plus an additional for garnish
2 thin slices of cucumber, plus an additional for garnish
Juice of 1 lime
2 ounces good quality blanco tequila
Club soda/sparkling water
In a mixing glass, muddle raspberries and cucumber. Add lime juice and tequila and shake thoroughly, without ice. Pour the contents into a rocks class filled with one large ice cube, top with club soda, a raspberry, and a slice of cucumber.
Manhattanites and Brooklynites, looking for an easier way to get your Whole Foods Market Cinco ingredients delivered straight to your door? Check out instacart to get your fiesta started.
|
{
"pile_set_name": "Pile-CC"
}
|
Gait analysis and electromyography in fixed- and mobile-bearing total knee replacement: a prospective, comparative study.
The theoretical superiority of mobile-bearing total knee arthroplasties (TKAs) has not yet been proven in clinical studies. The aim of the current study was to compare and to analyse in a patient population that had received either a fixed or a mobile TKA differences in gait analysis electromyography and clinical scores. In a prospective, randomized, patient- and observer-blinded clinical study, 33 patients with primary osteoarthritis of the knee were included. All patients received a Genesis II total knee replacement. Sixteen patients received a mobile and 17 a fixed-bearing cruciate retaining Genesis II TKA. Clinical and quality-of-life scores, electromyography and gait analysis were applied preoperatively and postoperatively with a follow-up of 24 months. In both groups, improvements from pre- to postoperative were detected. whereas the results of gait analysis and electromyography did not show any differences. The results from the clinical and the quality-of-life scores improved from pre- to postoperative, while the Knee Society Score showed a superiority of the mobile-bearing group (mean 159 ± 28) over the fixed-bearing group (mean 134 ± 41). No functional advantage of mobile over fixed-bearing TKA was detected, although the mobile-bearing group had better clinical results for which a reason could not be found. These results only apply to cruciate retaining mobile-bearing TKA with a bearing which allows both rotation and anteroposterior translation.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Molecular and cellular mechanisms of porphyrin photosensitization.
Mechanisms of porphyrin-sensitized photochemistry involving biomolecular substrates have been studied in a diverse array of in vitro and in vivo experimental protocols. Porphyrin-sensitized, singlet oxygen-mediated photooxidative damage has been implicated in peroxidation of cell membrane lipids; cross-linking of cell membrane and intracellular proteins; inhibition of cell membrane associated, cytosolic, mitochondrial and microsomal enzymes; disruption of intracellular organelles with release of inflammatory mediators and hydrolases; damage to DNA with retarded protein synthesis and delay in cell cycle; and activation of complement-generated inflammatory events leading to cell injury or death.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
---
abstract: |
Consider a random geometric graph $G(\chi_n, r_n)$, given by connecting two vertices of a Poisson point process $\chi_n$ of intensity $n$ on the unit torus whenever their distance is smaller than the parameter $r_n$. The model is conditioned on the rare event that the number of edges observed, $|E|$, is greater than $(1 + \delta)\mathbb{E}(|E|)$, for some fixed $\delta >0$. This article proves that upon conditioning, with high probability there exists a ball of diameter $r_n$ which contains a clique of at least $\sqrt{2 \delta \mathbb{E}(|E|)}(1 - \varepsilon)$ vertices, for any $\varepsilon
>0$. Intuitively, this region contains all the “excess” edges the graph is forced to contain by the conditioning event, up to lower order corrections. As a consequence of this result, we prove a large deviations principle for the upper tail of the edge count of the random geometric graph.
address:
- 'Department of Statistics Stanford UniversityEmail: '
- 'Department of Mathematics University of Geneva Email: '
author:
- Sourav Chatterjee
- Matan Harel
title: Localization in random geometric graphs with too many edges
---
[^1]
Introduction
============
The Random Geometric Graph is a simple stochastic model, first studied in [@Hafner] in 1972, for generating a graph: given the parameters $n$ and $r$, consider a Poisson point process of intensity $n$ on the unit torus, equipped with some norm $\| \cdot \|$, and declare an edge between any two vertices that are distance $\le r$ from each other.
Unlike the well-known Erdős–Rényi random graph, the random geometric graph’s definition leads to strong dependence between edges: if three vertices form a “V” shaped graph, they are far more likely to have the third edge of the triangle than if no assumption were made on the other edges, as a consequence of the triangle inequality.
Many properties of this graph model have been studied. The classic monograph of Mathew Penrose [@Penrose2] studies results pertaining to many graph-theoretical functions of random geometric graphs, including (but not limited to) laws of large numbers and central limit theorems for subgraph counts, independence number, and chromatic number, as well as many properties connected to the giant component. Many of the results presented in this monograph have been improved and generalized by Penrose and his coauthors in the 12 years since its initial publication. Besides this, there have been investigations into other probabilistic features, such as threshold functions for cover times and mixing times [@Avin] and a characterization of sharp thresholds for many monotone graph functions [@Goel]. This list is far from comprehensive, of course, and the random geometric graph is still an active research area.
The random geometric graph is also highly related to the random connection continuum percolation model. In that model, the vertex set is given by an (almost surely infinite) Poisson point process of fixed intensity on $\mathbb{R}^d$, and two points are connected with some probability that varies (and usually decreases) with their distance. In particular, the special case in which the radius of connection is deterministically fixed at 1 was the model that initiated the study of this kind of random geometry, in the seminal paper of Gilbert [@Gilbert]. The objects of interest in this model are the existence of an infinite connected component, as well as the behavior of the subset of $\mathbb{R}^d$ that is at distance at most 1 from one of the vertices of the graph (the so-called “Poisson blob”). Continuum percolation is treated in detail in a book-length monograph by Meester and Roy [@Meester], as well as in the more general work of Grimmett [@Grimmett].
Most of the work done on random geometric graphs is concerned with either the behavior of a typical graph — the graph we are likely to see for a given $r$ as $n$ goes to infinity — or typical deviations from that behavior, i.e. central limit theorems. In this paper, we are concerned with the behavior of the model conditioned on a rare event. Specifically, we are concerned with the random geometric graph conditioned on having more edges than is expected (a formal description will follow). The large deviation regime of the upper tail of any subgraph count of the random geometric graph is not well studied, though some bounds are available: Janson [@Janson] has studied concentration inequalities for $U$-statistics, a general class of statistics which includes the subgraph counts we are interested in. These upper bounds work in very general settings, but are not tight, even up to constants in the exponent. Large deviation principles have been proven for functionals of random point processes in which the contribution of any particular vertex is uniformly bounded [@Schreiber], but no such bound is known for functionals with possibly large influence, such as the edge count of the graph.
As motivation for this detailed study, we consider the problem in a more familiar context: the “infamous upper tail" [@Janson2] of the triangle count $T$ in the Erdős–Rényi random graph, $G(n,p)$. After many years of development of increasingly strong bounds, the first breakthrough was made by Kim and Vu [@Vu] and Janson [@Janson3] independently, who proved that, for any $\delta >0$ and whenever $p \gg (\log n)/n$, $$\exp[ - c(\delta) n^2 p^2 \log(1/p)] \leq \mathbb{P}[T > (1 + \delta)\mathbb{E}[T]] \leq \exp[- C(\delta) n^2 p^2] \, ,$$ where $c(\cdot)$ and $C(\cdot)$ depend only on $\delta$. Recently, there has been renewed interest in these type of tail estimates. In 2010, Chatterjee [@Chatterjee2] and Demarco and Kahn [@Demarco] (in independent works) established the correct order of the upper tail of triangles and other small cliques by adding the missing logarithmic term to the upper bound, without providing good control of the leading-order constants. The recent work of Chatterjee and Dembo [@Chatterjee3] on nonlinear large deviations allowed Lubetzky and Zhao [@Lubetzky] to calculate both the order and the leading-order constant for the upper tail question in a certain regime of sparse Erdős–Rényi random graphs. Unfortunately, the technique is based on the analysis of polynomials in Bernoulli random variables, and thus ill-suited to the problem we are studying here.
In this work, we use the structure of the random geometric graph to evaluate the upper tail large deviation rate function. In addition, we provide a “structure theorem" to describe the graph-theoretical structure of the model conditioned on having too many edges. Specifically, we show that such a conditional model exhibits [*localization*]{}. Heuristically, this phenomenon can be described as a scenario in which a small number of vertices will contribute almost all the extra edges that we require the graph to exhibit, while the edge count of the “bulk” of the graph of the graph will remain largely unchanged, in some weak sense. Furthermore, we will show that the geometry of the localized region has the shape of a ball in the given norm (we will make these two statements more precise at the end of the next section). Outside of the aforementioned work of Lubetzky and Zhao [@Lubetzky] in the Erdős–Rényi model, which is only known for a limited range of $p$’s, this work is the first (as far as the authors are aware) to establish that the large deviation regime of a subgraph count is (weakly) equivalent to planting a combinatorial structure in the usual, unconditional graph.
The fact that large deviation regimes may be dominated by configurations with a small number of very large contributions was known relatively early in the history of large deviation theory: a survey by Nagaev [@Nagaev], summarizing a series of papers written in the Soviet Union in the 1960’s and 70’s, includes this observation for sums of i.i.d. random variables with stretched-exponential tails as a corollary. In our context, the natural combinatorial structure for creating many edges with a small number of edges is a “giant clique." The clique number, the (typical) size of the largest clique of the random geometric graph, falls under the general class scan statistics, has been shown to focus on two values with high probability for certain values of the threshold parameter $r$ (see [@Penrose4], [@Muller]); however, these works do not explore the large deviation regime. Our work uses techniques from large deviations, concentration inequalities, convex analysis, and geometric measure theory. A key component in the proof is a technique for proving localization that has previously appeared in [@Talagrand1] and [@Chatterjee].
Main result
===========
Let $\chi_n$ be a Poisson Point Process of intensity $n$ on the $d$-dimensional unit torus $\mathbb{T}^d = [0,1]^d$. Let $N := |\chi_n|$. Recall that $N$ is a Poisson random variable with mean $n$, and conditional on $N$, $\chi_n$ is just a set of $N$ points chosen independently and uniformly at random. Let $r_n$ be a positive sequence that decreases to $0$ as $n\to \infty$, and $\| \cdot \|$ be some norm on $\mathbb{R}^d$ that induces a translation-invariant metric on $\mathbb{T}^d$. We define the random geometric graph $G(\chi_n,r_n) : = (V,E)$, where $V = \chi_n = \{v_1,\ldots, v_N\}$, enumerated arbitrarily, and $E$ is the set of unordered pairs $\{i,j\}$ such that $\| v_i - v_j\| \leq r_n$. Figure 1 shows a particular instance of $G(\chi_{150},0.1)$.
Letting $1_{i,j}$ be the indicator that there is an edge between $v_i$ and $v_j$, we can calculate the expected value of $|E|$, the number of edges in the graph: $$\begin{aligned}
\mathbb{E}(|E|) &= \mathbb{E}\left(\displaystyle \sum_{1\le i<j\le N} 1_{i,j} \right) = \mathbb{E}\left[ {N \choose 2} \mathbb{E}(1_{1,2} \mid N) \right]\\
&= \frac{n^2}{2} \mathbb{P}(\|v_1-v_2\|\le r_n).\end{aligned}$$ By the symmetry of the torus, $\mathbb{P}(\|v_1-v_2\|\le r_n)$ is simply $\nu r_n^d$, where $\nu$ is the volume of the unit ball of the given norm, $r_n$. Thus, $$\mathbb{E}(|E|) = \frac{n^2 \nu r_n^d}{2} =: \mu_n.$$ For the rest of the article, we suppose the existence of a fixed constant $\delta^* >0$ such that, for all sufficiently large $n$, $$\label{rcond}
n^{ (\delta^*-2)/d } \leq r_n \leq n^{- \delta^*/d}$$ The lower bound ensures that the expected number of edges is polynomially large in $n$; the upper bound excludes the possibility of $r_n$ vanishing as the inverse of a polynomial in $\log n$ (for example), but is not much stronger than the initial requirement of $r_n$ vanishing in as $n$ grows. We define a parameter $p$ as $$\label{pdef}
p : = \displaystyle \lim_{n \rightarrow \infty} \frac{ \log \mu_n}{\log n}\, ,$$ implicitly assuming that the limit exists; we will continue to assume this for the rest of the paper.
This ensures that $\mu_n = f(n) n^p$, where $f$ varies more slowly than any polynomial or rational function in $n$. Notice that $p$ depends only on the tail behavior of $r$, and is completely independent of $n$. Furthermore, $$\label{pbounds}
\delta^* \leq p \leq 2 - \delta^*\, ,$$ thanks to .
The following theorem is the main result of the paper:
\[thm1\] Let $G(\chi_n,r_n)$ be a random geometric graph model on the d-dimensional torus with respect to some norm $\| \cdot \|$ and a threshold parameter that satisfies $n^{ (\delta^*-2)/d } \leq r_n \leq n^{- \delta^*/d}$ for some fixed $ \delta^*>0$. Let $\tau_n = \nu (r_n/2)^d$, where $\nu$ is the volume of the unit ball under the norm $\|\cdot \|$. (That is, $\tau_n$ is the volume of the ball of diameter $r_n$.) Fix $\delta >0$ and $\varepsilon > 0$. Let $F_n$ be the event that the following happen: (a) There is a ball $A$ of diameter $r_n$ such that any convex set $S\subseteq A$ with $\lambda (S) >
(\varepsilon/16) \tau_n$ (where $\lambda$ denotes Lebesgue measure) satisfies $$\left|\frac{|\chi_n(S)|}{\sqrt{2 \delta \mu_n}} - \frac{\lambda(S)}{\tau_n}\right| < \varepsilon\, ,$$ and (b) for any convex set $S\subseteq A^c$ with $\lambda(S) > (\varepsilon/16)\tau_n$, such that $S$ lies in some ball of diameter $r_n$, $$\frac{|\chi_n(S)|}{\sqrt{2 \delta \mu_n}} < \varepsilon\,\frac{\lambda(S)}{\tau_n}\, .$$ Then the conditional probability of the event $F_n$ given that $|E| \geq (1+\delta) \mu_n$ tends to $1$ as $n\to \infty$.
The convexity requirement is probably not optimal, but forces the set $S$ to be sufficiently “nice” to preclude sets which are either sparse but of large measure (such as generalized Cantor sets) or have boundaries that take up a large amount of space. We also force $S$ to be large to make sure it cannot pick up the lower order uneveness in the conditional process. Probing the Poisson Point Process with such sets is sufficient to show that graph will appear uniform, up to sets of measure equal to arbitrarily small multiples of $\tau_n$.
As a consequence of Theorem \[thm1\], we prove that the upper tail of the edge count of random geometric graphs satisfies a large deviation principle. Recall that $(|E|-\mu_n)/\mu_n$ satisfies an [*upper tail*]{} large deviation principle with speed $f(n)$ and rate function $I(x)$ if, for any closed set $F \subset (0,\infty)$, $$\displaystyle \limsup_{n\to \infty}\frac{1}{f(n)} \log \mathbb{P}\left(\frac{|E| - \mu_n}{\mu_n} \in F \right) \leq -\displaystyle \inf_{x \in F} I(x)\,
,$$ and for any open set $G \subset [0,\infty)$, $$\displaystyle \liminf_{n\to\infty} \frac{1}{f(n)} \log \mathbb{P}\left(\frac{|E| - \mu_n}{\mu_n} \in G \right) \geq - \displaystyle \inf_{x \in G} I(x)\,
.$$
\[thmLDP\] Let $G(\chi_n,r_n)$ be a random geometric graph model on the d-dimensional torus, with the same assumptions as in Theorem \[thm1\]. Define $$I(x) := \left(\frac{2-p}{2}\right) \sqrt{2 x}\, ,$$ where $p$ is defined as in . Then the random variable $(|E|-\mu_n)/\mu_n$ satisfies an upper tail large deviation principle with speed $\sqrt{\mu_n} \log n$ and rate function $I(x)$.
Note that this rate function, like all other claims in this paper, is only valid for events in which the number of edges exceeds its mean. The lower tail of the edge count is likely to satisfy Poisson-like statistics, and hence the speed of its large deviation probability is expected to be of order $\mu_n$, not $\sqrt{\mu_n} \log n$.
Before we go on, we’d like to comment on the heuristic formulation we presented earlier. First, since the rate function of \[thmLDP\] is strictly increasing, we know that, conditional on $\{|E| > (1 + \delta) \mu_n\}$, the event $$\left\{(1 + \delta') \mu_n > |E| > (1 + \delta) \mu_n\right\}$$ occurs with high probability for any $\delta' > \delta$ and $n$ sufficiently large. If we take $S = A$ in the first stipulation of \[thm1\], we see that the ball $A$ of diameter $r_n$ includes at least $\delta \mu_n(1 - \varepsilon)$ edges of the random geometric graph. Since $\varepsilon$ and $\delta' -
\delta$ are arbitrarily close to zero, we recover our original claim that ‘almost all extra edges are between points in $A$’. In terms of the geometry, we see that there are approximately $\sqrt{2 \delta \mu_n}$ vertices that are more or less equi-distributed inside $A$. Outside of this set, we would like to say that there are no other large cliques; unfortunately, \[thm1\] does not provide this result. Instead, we can only be sure that every other clique outside of the “exceptional" set $A$ has $o(\sqrt{\mu_n})$ vertices. We conjecture that the size of the largest clique disjoint from $A$ is of the same order as the size of the largest clique of an [*unconditional*]{} random geometric graph with the same values of $n$ and $r_n$. A combination of this conjecture and elementary estimates on Poisson Point Processes should be sufficient to establish a (suitably formalized) version of the claim ‘the remaining portion of the graph remains unchanged.’
The $s$-Graded Model
====================
Henceforth in the manuscript, we will suppress the subscript $n$ and write $\chi$, $\mu$, $\tau$ and $r$ instead of $\chi_n$, $\mu_n$, $\tau_n$ and $r_n$.
We now present an approximation of the random geometric model which allows us to replace the Poisson point process with a sequence of independent Poisson random variables. To do this, we first discretize space, and then produce a metric on the resulting “cells” that approximates the norm $\| \cdot \|$ on the unit torus. We call this the $s$-graded model.
Fix an integer $s$, and define $$m := \lfloor s/r \rfloor,$$ so that $$\frac{s}{r} - 1 \leq m \leq \frac{s}{r}.$$ Let $T = \{1,2, \dots m\}^d$. Pick $I = (i_1, i_2, \dots, i_d) \in T$, and define $$A_I = \left[\frac{i_1-1}{m},\frac{i_1}{m}\right] \times \dots \times \left[\frac{i_d-1}{m},\frac{i_d}{m}\right].$$ The $A_I$’s partition the unit torus into $m^d$ cubes (ignoring sets of measure $0$), each of volume $1/m^d$, and therefore, $X_I = |\chi(A_I)|$ is a Poisson random variable of mean $$\mathcal{D} := \frac{n}{m^d}.$$ As a function of $n$, $m^d$ grows as $n^{2-p}$, up to a possible multiplicative factor that varies more slowly than any polynomial or rational function in $n$ (recall that $s$ is fixed); this implies that $\mathcal{D}$ grows as $n^{p-1}$. Note that $s$ is exactly the minimal number of cells that must be traversed in any one axial direction to ensure that we exceed the distance $r$. This parameter takes the place of $r$ in the $s$-graded model.
We now define a metric on $T$, induced by the norm on torus: $$d(I,J) = \inf_{x \in A_I^o,\, y \in A_J^o} \left\lceil m \| x - y \|\right\rceil$$ where the circles indicate the interiors of the sets. Note that the distance is always an integer. Moreover, $d(I,J) = z$ if $z$ is the smallest integer such that some point in $A_I^o$ and some point in $A_J^o$ are less than $z$ away, measured in units of $1/m$, the side length of the cubes. We force the points to be in the interior to prevent “trivialities”, such as two adjacent cells being distance $0$, since they share a boundary.
We are now ready to define the $s$-graded random geometric graph. Let $G_s(\chi,r) = (V,E_s)$ have the same vertex set as the original graph. For each vertex $v$, let $I_v$ be the index in $T$ such that $v \in A_{I_v}$; there is ambiguity on the boundary of the $A_I$’s, but that set has Lebesgue measure 0, and therefore it has no vertices of $\chi$, almost surely. We say $(v,w) \in E_s$ whenever $d(I_v, I_w) \leq s$. Essentially, the $s$-graded model allows every point to wander inside a cubical “cage” of side-length $1/m$, and connects any two points that might be connected after we allow this mobility. In this framework, it is clear that $E_s$ becomes smaller as $s$ decreases. In fact, for sufficiently large $s$, $E_s$ is identical to $E$; unfortunately, this $s$ will depend on $n$. Due to uniformity issues, we will instead fix $s$ and later show that, even when $s$ is finite but large, the approximation is good.
The major benefit of the $s$-graded model is that its edge count is very simple to express in terms of $X_I$, the number of points in each $A_I$: $$\begin{aligned}
|E_s| &= \displaystyle \sum_{I \in T} \biggl[{X_I \choose 2} + \frac{1}{2} \sum_{J\, : \, 0 < d(I,J) \leq s} X_I X_J \biggr]\\
&= \frac{1}{2}\sum_{I \in T} X_I \left[\biggl(\displaystyle \sum_{J\, : \, d(I,J) \leq s} X_J\biggr) -1\right].\end{aligned}$$ This random variable is defined in terms of i.i.d. random variables, which eases the analysis greatly. Furthermore, the ambient geometry of the torus is encoded completely by the metric $d$. Finally, each $X_I$ only appears in finitely many terms in this expression. The “finite range" nature of the representation will play a major role in the proof presented.
We quantify this fact as follows: for any $I \in T$, let $N_I := \{J: d(I,J) \leq s\}$. Thanks to translation invariance of $d$, the cardinality of this set is independent of the choice of $I$. Using this parameter, we can compute the expected number of edges in the $s$-graded random geometric graph easily: $$\begin{aligned}
\tilde{\mu}_s &:= \mathbb{E}(|E_s|)\label{musdef}\\
&= \displaystyle \sum_{I \in T} \mathbb{E}\biggl[{X_I \choose 2}\biggr] + \frac{1}{2} \sum_{J\, : \, 0 < d(I,J) \leq s} \mathbb{E}(X_I)
\mathbb{E}(X_J)\nonumber \\
&= \frac{|N_I| m^d \mathcal{D}^2}{2} = \frac{|N_I|n^2}{2m^d}\, . \nonumber\end{aligned}$$ As before, we are interested in conditioning the $s$-graded model on the event $\{|E_s| > (1 +\tilde{\delta}) \mu_s\}$. Again, the appropriate geometric notion is that of the largest set of diameter $s$. We call a set of indices a $\textit{maximal clique set}$ if it is a subset of $T$ with diameter $s$ that achieves the maximal cardinality of all such sets. We define $$\label{tausdef}
\tilde{\tau}_s := \max \{ |\mathfrak{I}| : \mathfrak{I} \subset T, \text{diam}(\mathfrak{I}) \leq s \} \, ,$$ i.e. $\tau_s$ is the cardinality of a maximal clique set.
We can now state the equivalent to Theorem \[thm1\] for the $s$-graded model:
\[thm2\] Fix an integer $s$ and let $G_s(\chi,r)$ be an $s$-graded random geometric graph, with $ n^{(2-\delta^*)/d} \leq r \leq n^{-\delta^*/d}$ for some fixed $\delta^*>0$. Fix $\tilde{\delta}>0$ and $\tilde{\varepsilon}>0$. Let $F$ be the event that the following happen: (a) There is a maximal clique set $\mathfrak{P}$ such that for all $I\in \mathfrak{P}$, $$\biggl|\frac{ \tilde{\tau}_s X_I}{(2 \tilde{\delta} \tilde{\mu}_s)^{1/2}} - 1 \biggr| < \tilde{\varepsilon}\, ,$$ and (b) for all $J\in \mathfrak{P}^c$, $$\frac{\tilde{\tau}_s X_J}{(2 \tilde{\delta} \tilde{\mu}_s)^{1/2}} \leq \tilde{\varepsilon}\, .$$ Then the conditional probability of $F$ given $|E_s| \geq (1+\tilde{\delta}) \tilde{\mu}_s$ tends to $1$ as $n\to \infty$.
Essentially, Theorem \[thm2\] produces a maximal clique set, and whose entries sum up to $(2 \tilde{\delta} \tilde{\mu}_s)^{1/2}$, up to lower order corrections. This creates all the ‘extra edges’ in the graph. As in Theorem \[thm1\], this set will contain almost all the edges not expected to be the in unconditional $s$-graded model, and the rest of graph will look unaltered, at least in the scale of $\sqrt{\tilde{\mu}_s}$. This theorem will also be instrumental in proving \[thmLDP\]; that proof will appear in Section 7.
Outline of the Proof
====================
Before embarking on a proper proof, we sketch the main ideas required. The first step is to show that the $s$-graded model is, in fact, a good approximation for the random geometric graph. To do so, we first show that we can approximate any convex subset $S$ of a ball of diameter $r$ from both the inside and the outside by a union of $A_I$’s. Next, we use the classical isodiametric inequality to show that the $A_I$’s associated with a maximal clique set approximate a ball of diameter $r$, in the sense of the Hausdorff metric. Armed with these facts, showing that Theorem \[thm2\] implies Theorem \[thm1\] is a matter of careful “epsilontics.”
We then turn to directly analyzing the $s$-graded model, conditioned on the event $$L := \{|E_s| \geq (1+\tilde{\delta}) \tilde{\mu}_s\}\, .$$ For notational convenience, let $q = \sqrt{2 \tilde{\delta} \tilde{\mu}_s}$, $w = \tilde{\tau}_s \mathcal{D}$. Define $$\mathfrak{I} = \{ I \in T : X_I > \max\{\mathcal{D}n^a, n^a\} \}\, ,$$ where $a$ is a small positive number. This is the set of “large” indices, in the sense that their associated $X_I$’s exceed their expected value by a fixed polynomial factor in $n$. Furthermore, define $$Y_I := X_I \left(\log (X_I /\mathcal{D}) -1\right) + \mathcal{D} \, ,$$ and $$Q(\mathfrak{I}) := \frac{2}{q^2}\biggl( \displaystyle \sum_{I \in \mathfrak{I}} {X_I \choose 2} + \frac{1}{2} \displaystyle \sum_{J \in N_I \cap
\mathfrak{I}} X_I X_J\biggr) \,.$$ The former quantity is an appropriately chosen convex function of the $X_I$’s, while the latter is a scaled version of the number of edges with both endpoints in the $A_I$’s associated with $\mathfrak{I}$. Consider the event $$\left\{Q(\mathfrak{I}) > 1 - \frac{\xi}{\log n}\right\} \cap \left\{\frac{1}{q} \displaystyle \sum_{I \in \mathfrak{I}} Y_I \leq \log(q/w) -1 +
\xi\right\}\, ,$$ where $\xi$ is an arbitrary positive constant. By comparing a lower bound on the probability of $L$ to an upper bound on the probability of the [*complement*]{} of the above event, we can show that this event occurs with high probability in the $s$-graded model conditional on $L$.
We now have a set of indices $\mathfrak{I}$ which satisfies both a quadratic lower bound and a convex upper bound with high probability. From here on, the analysis is completely deterministic, characterizing configurations that satisfy the simultaneous inequalities above. We wish to look at the largest elements of $\mathfrak{I}$. Specifically, we wish to take the smallest set that includes roughly $q$ vertices of the Poisson point process. To make this precise, define $$V(A) := \frac{1}{q} \displaystyle \sum_{I \in A} X_I \, ,$$ for any subset of indices $A$. Now, order the elements of $\mathfrak{I}$ by size, so that $X_1$ is the largest, $X_2$ is the second largest, etc. Letting $\mathfrak{T}_k = \{X_1, X_2, \dots X_k\}$, define $$\mathfrak{T} := \text{ The first } \mathfrak{T}_k \text{ such that } V(\mathfrak{T}_k) > 1 - \frac{2 \xi}{\log n}\, .$$ Careful use of minimality and Jensen’s inequality can establish that $$V(\mathfrak{T}) \leq 1 + \phi(\mathfrak{T}) \, ,$$ where $\phi(\mathfrak{T})$ is bounded above by $\xi/3$. More importantly, moving from $\mathfrak{I}$ to $\mathfrak{T}$ does not force us to discard too many edges; formally, $$Q(\mathfrak{T}) \geq 1 - \psi(\mathfrak{T}) \, ,$$ with $\psi(\mathfrak{T})$ bounded above by $\xi$.
We now define $$\mathfrak{P} := \left\{ I \in \mathfrak{T} : X_I > \frac{\xi^{1/4} q}{\tilde{\tau}_s} \right\}\, .$$ If $\mathfrak{I}$ was the set of large indices, $\mathfrak{P}$ is the set of very large indices — that is, those whose associated $X_I$’s are commensurate with $q$. We can show that this set cannot have diameter that strictly exceeds $s$ without contradicting either the lower bound $Q(\mathfrak{T})$ or the upper bound on $V(\mathfrak{T})$. Furthermore, technical estimates ensure that $|\mathfrak{P}| \geq \tilde{\tau}_s$. As a set of diameter at most $s$ and cardinality at least $\tilde{\tau}_s$, $\mathfrak{P}$ must be a maximal clique set, by definition. Moreover, a quantitative version of Jensen’s inequality allows us to claim that all the $X_I$’s associated with this set are roughly equal. Thus, $\mathfrak{P}$ is the maximal clique set described in Theorem \[thm2\]. Finally, we can show that its complement is made up on $X_I$’s that vanish in comparison to $q$. This completes the proof.
In the last section, we prove the large deviation principle. We use the first stipulation of \[thm1\] and the $s$-graded model to compute the upper bound. The lower bound is derived directly from the Poisson Point Process.
The $s$-Graded Model Approximates the Random Geometric Graph
============================================================
Before proving Theorem \[thm2\], we show that it implies Theorem \[thm1\]. To do so, we first define three operations to go between subsets of $T$, the natural objects in the $s$-graded model, and subsets of $[0,1]^d$. For $\mathfrak{B} \subset T$, let $$\mathfrak{U}(\mathfrak{B}) := \displaystyle \bigcup_{I \in \mathfrak{B}} A_I$$ the set associated with the index set $\mathfrak{A}$. In the other direction, we cannot be as exact; for $K \subset [0,1]^d$, let $\mathfrak{R}(K)$ and $\mathfrak{O}(K)$ be the outer and inner hulls of $K$, defined as the minimal and maximal subsets of $T$ such that $$\mathfrak{U}\left(\mathfrak{R}(K)\right) \subset K \subset \mathfrak{U}\left(\mathfrak{O}(K)\right)$$ With these operators in place, we show that the $s$-graded parameters are good approximations of their respective properties in the usual random geometric graph:
\[lmm3\] Let $G(\chi,r) = (V,E)$ be the usual random geometric graph, and $G_s(\chi,r) = (V,E_s)$ be the $s$-graded model. As before, $\mu = \mathbb{E}(|E|)$, $\tilde{\mu}_s = \mathbb{E}(|E_s|)$ and $\tau = \nu (r/2)^d$. Let $\tilde{\tau}_s$ be the cardinality of a maximal clique set (as defined in ). Then $E\subseteq E_s$, and there exist constants $C$ and $s_0$, depending only on the dimension and the chosen norm of the torus, such that if $s\ge s_0$ then $$\mu \leq \tilde{\mu}_s \leq \mu \left(1 + \frac{C}{s}\right)$$ and $$m^d \tau \leq \tilde{\tau}_s \leq m^d \tau \left(1 + \frac{C}{s}\right)$$ Furthermore, both $|N_I|$ and $\tilde{\tau}_s$ are uniformly bounded in $n$.
Pick an arbitrary $I$ and consider $\mathfrak{U}(N_I)$. By definition of $d$ and $s$, this set includes a ball of radius $r$ around $\textit{any}$ point in $A_I$. Therefore, any pair $(v,w) \in E$ must also be in $E_s$, giving the first stipulation. Since this inclusion holds for any configuration of the underlying Poisson Point process, this also gives $\mu \leq \tilde{\mu}_s$.
Now, let $\rho$ be the diameter of the unit cube under the norm $\| \cdot \|$. Then, for any $x \in A_I$ and $y \in \mathfrak{U}(N_I)$, $$\| x - y\| \leq \frac{rs}{s-r} + \frac{2 \rho}{m}$$ Therefore, $\mathfrak{U}(N_I)$ is contained in a ball of radius $r + 3\rho/m$ around any point in $A_I$, which implies that $$|N_I| = m^d \lambda\left(\mathfrak{U}(N_I)\right) \leq \nu m^d r^d \left(\frac{s}{s-r} + \frac{3 \rho}{rm}\right)^d \leq \nu m^d r^d \left(1 + \frac{3r +
6 d \rho}{s-r}\right)$$ where the final inequality follows because $(1 +x)^d \leq 1 + (2d)x$ for all sufficiently small $x$, and the fact that $rm \geq s-r$ by definition of $m$. Substituting this into the definition of $\tilde{\mu}_s$ produces the desired inequality on $\tilde{\mu}_s$ (assuming, without loss, that $r\le \rho$).
To show the lower bound on $\tilde{\tau}_s$, let $B$ be a ball of radius $r/2$. $\mathfrak{U}(B)$ will have diameter at most $s$, by definition of the metric, and its measure will be at least $\tau$. Converting this to cardinality gives the desired lower bound. For the upper bound, let $\mathfrak{W}$ be a set of indices such that $$\lambda \left(\mathfrak{U}(\mathfrak{W})\right) \geq \tau \left(1 + \frac{ C}{s} \right)$$ Applying the isodiametric inequality for finite dimensional normed spaces [@Burago p. 93] and choosing $C$ and $s_0$ sufficiently large gives $$\text{diam}(\mathfrak{U}(\mathfrak{W})) \geq r \left(1 + \frac{ C}{s} \right)^{1/d} \geq r + \frac{4 \rho}{m}$$ This implies that the diameter of $\mathfrak{W}$ is at least $s+1$ in the $d$ metric, and therefore the set cannot be a maximal clique set. Translating the measure bound to a cardinality bound on the size $\mathfrak{W}$ gives the desired upper bound on $\mathfrak{W}$.
Finally, $m^d \leq s^d /r^d$, giving the uniform upper bound on $|N_I|$. Since $\tilde{\tau}_s \leq |N_I|$, the final claim of the lemma follows.
As a simple corollary to the lemma, we see that the hypothesis of Theorem \[thm1\] implies the hypothesis of Theorem \[thm2\], as long as $\tilde{\delta} < \delta$ and $s$ is sufficiently large:
\[DeltaCor\] For any $\tilde{\delta} <\delta$, there exists $s_0$ depending only on $\delta$, the chosen norm, and the dimension, such that whenever $s\ge s_0$, $|E| >
(1+\delta) \mu$ implies $|E_s| > (1+\tilde{\delta})\tilde{\mu}_s$.
Pick $s$ large enough to ensure $(1 +\tilde{\delta}) (1 + C/s) \leq 1 + \delta$. Lemma \[lmm3\] guarantees that $|E_s| \geq |E|$ and $$|E| \geq (1+\delta) \mu \geq (1 +\tilde{\delta}) (1 + C/s) \mu \geq (1 +\tilde{\delta}) \tilde{\mu}_s$$ This proves the corollary.
Next, we get some quantitative bounds on the geometric approximation of certain subsets of $[0,1]^d$ by the $A_I$’s. Let $S$ be a convex subset of a ball of diameter $r$. Assume that $$z(S) := \frac{\lambda(S)}{\tau} \geq \varepsilon/16\, .$$ Recall that $\tau$ is the volume of a ball of diameter $r$. We rescale Lebesgue measure to bypass factors of $\tau$ throughout this section. The first goal is to show that the inner and outer hulls of $S$ approximate $S$ uniformly in the choice of $S$, in the sense of the measure of the sets. Formally, we wish to prove the following proposition:
\[FatteningProp\] Fix $\varepsilon >0$, and let $S$ be a convex subset of a ball of radius $r$ with $z(S) \geq \varepsilon/16$. Then, there exists an $s_0$ that depends only on $\varepsilon$, the dimension, and the choice of norm, such that whenever $s > s_0$, the inner hull $\mathfrak{U}(S)$ satisfies $$z(S) -\frac{\varepsilon^2}{64} \leq z\big(\mathfrak{U}(\mathfrak{R}(S)) \big) \leq z(S)$$ Similarly, the corresponding inequality $$z(S) \leq z\big(\mathfrak{U}(\mathfrak{O}(S)) \big) \leq z(S) + \frac{\varepsilon^2}{64}$$ hold for the outer hull of $S$.
The essential element of this proposition is that $s_0$ is completely independent of the precise choice of $S$, as long as $S$ satisfies a diameter upper bound and measure lower bound. For a single choice of $S$, the continuity of Lebesgue measure is enough to establish these inequalities; however, the choice of $s_0$ given by this construction would depend on the geometry of $S$, which is insufficient for our purposes.
Instead of proving this statement directly, we consider this minor modification of the proposition:
\[BoundaryFattening\] Let $S$ be a convex set as in \[FatteningProp\]. Define $$(\partial S)_{l} := \left\{y : \| \partial S -y \| \leq l\right\}\, .$$ Fix $\varepsilon >0$. Then, there exists an $s_0$, depending on $\varepsilon$, the dimension, and the norm, such that $s>s_0$ implies $$z\left((\partial S)_{\rho/m}\right) \leq \frac{\varepsilon^2}{64}$$
This lemma implies \[FatteningProp\], since $$S \subset \mathfrak{U}\left(\mathfrak{R}(S)\right) \cup (\partial S)_{\rho/m}$$ and $$\mathfrak{U}(\mathfrak{O}(S)) \subset S \cup (\partial S)_{\rho/m} \, .$$ Evaluating $z(\cdot)$ of both sides and using subadditivity gives the two nontrivial bounds required in \[FatteningProp\]; the other two follow by inclusion.
Heuristically, the volume of $(\partial S)_{\rho/m}$ should be commensurate with the product of $\rho/m$ with the surface area of $S$. Since $S$ has diameter at most $r$ and is a convex set (and therefore its boundary cannot be too convoluted), this surface area ought to be bounded above by $Cr^{d-1}$, and therefore $z[(\partial S)_{\rho/m}]$ will be small. To formalize this loose heuristic, we turn to the tools of geometric measure theory. Although this approach is standard in that field, we include a detailed proof for completeness. The analysis is slightly technical, and breaks down to three parts: First, we state the coarea formula for Lipshitz functions, defining all relevant terms. Next, we establish that the function $f(x) = \|\partial S-x \|$ is Lipshitz, and use the coarea formula to express the measure of $(\partial S)_{\rho/m}$ as an integral of the $(d-1)$ dimensional Hausdorff measures of well-behaved sets. Finally, we show that, thanks to convexity of $S$, the measures in the integrand can be uniformly bounded by $Cr^{d-1}$; a computation completes the proof.
Consider a function $f:\mathbb{R}^d \rightarrow \mathbb{R}$ that is Lipschitz with respect to the Euclidean distance, and a Borel set $A$ . The Euclidean coarea formula [@Federer pg. 248] states that, with the functions as above, $$\displaystyle \int_{A} \|Df(x)\|_2 dx = \displaystyle \int_{-\infty}^{\infty} H^{d-1}\left(A \cap f^{-1}(y) \right) dy\, ,$$ where $\| \cdot \|_2$ is the Euclidean norm, and $H^{d-1}$ is the Hausdorff measure on the surface (effectively the surface area) and $Df$ is the gradient of $f$, which exists since $f$ is almost everywhere differentiable. To formally define this, pick a set $A$ and let $U_\delta(A)$ be the set of all coverings $\{U_i\}$ of $A$, where $U_i$ have diameter at most $\delta$. By definition, the Hausdorff measure of $A$ is $$H^{d-1}(A) = C_{d-1} \displaystyle \lim_{\delta \downarrow 0} \displaystyle \inf_{\{U_i\} \in U_\delta(A)} \displaystyle \sum [\text{diam } (U_i)]^{d-1}
\, ,$$ where $C_{d-1}$ is some constant depending on dimension, used to normalize the measure appropriately to be compatible with Lebesgue measure. This limit is well defined [@Federer pg. 170], but may be $0$ or infinity for a general $A$; in fact, it is infinite for any set with positive $d$-dimensional Lebesgue measure.
Now, consider the function $f(x) = \|\partial S - x\|$, where $\|B-x\|$ is shorthand for $\inf_{b\in B} \|b-x\|$ for any set $B$. First, we recall the classical fact that all norms are equivalent in finite dimensional space, i.e. there exists two positive constants $c$ and $C$ such that, for all $x$ and $y$, $$c \|x -y\| \leq \|x -y\|_2 \leq C \|x- y\|\, .$$ To see that $f$ is Lipschitz with respect to the Euclidean norm, pick two points $x$ and $y$, and let $a \in \partial S$ be the point such that $f(x) =
\|x - a\|$ (this point exists because $\partial S$ is closed). Then $$f(y) - f(x) \leq \|y - a\| - \|x - a\| \leq \|x-y\| \leq C \|x-y\|_2\, .$$ Similarly, $f(x)-f(y)\le C\|x-y\|_2$. Thus, $f$ is differentiable almost everywhere. Pick an $x$ where the function is differentiable, and let $a$ be as before. Then, for any $t \in
(0,1)$, $$f(x + t(a-x)) \le f(x) - t \|a-x\|\, ,$$ by the properties of norms. Subtracting $f(x)$ from both sides, dividing by $t$ and letting $t\to 0$, we get $$\langle a-x, Df\rangle \le - \|a-x\|\ \, ,$$ where $\langle \cdot, \cdot \rangle$ is the Euclidean inner product (or in this case, the directional derivative). Applying the Cauchy-Schwarz inequality, we conclude that $$\|a-x\| \leq \|a-x\|_2\|Df\|_2\, ,$$ which implies, by the equivalence of norms, that $$\|Df\|_2 \geq \frac{\|a-x\|}{\|a-x\|_2} \geq c\, .$$ Substituting this into the coarea formula, we see that $$c \lambda(A) \leq \displaystyle \int_{-\infty}^{\infty} H^{d-1}\left(A \cap f^{-1}(y)\right) dy\, ,$$ for any choice of $A$.
Letting $A = \{ x : f(x) \leq \rho/m\}$, we deduce that $$\lambda\left[(\partial S)_{\rho/m}\right] \leq C \displaystyle \int_0^{\rho/m} H^{d-1}\left(\{y : \|\partial S - y\| = z \} \right) dz \, ,$$ where $C$ is some (possibly different) universal constant. Note that, for sufficiently small $z$ and $T$ convex, the set $\{y : \|\partial T -y\| = z \}$ has two parts: one inside $T$ and the other outside of it. Luckily, the external one is the boundary of $T_z$, which is convex as it is the affine sum of $T$ and a ball of radius $z$. The internal one is boundary of the set $$T^{(z)} := \{x : x \in T, \|\partial T - x\| \geq z\}\, .$$ This set is also convex; it if it weren’t, we could find $x,y \in T^{(z)}$ such that $w = t x + (1-t) y \not \in T^{(z)}$ for some $t \in (0,1)$. Let $v$ be the minimal length vector such that $w + v \in \partial T$. Then, since $\|v\| < z$ by definition, we can find an $\varepsilon$ sufficiently small such that $w + (1+\varepsilon) v \not \in T$, while $x +(1 + \varepsilon)v$, $y +(1 + \varepsilon)v$ are both in $T$, contradicting convexity of $T$. Therefore, controlling the “fattening” of $\partial S$ reduces to controlling the boundaries of convex sets.
Suppose $T$ is a convex subset of a Euclidean ball $B$ centered around some $x$ in the interior of $T$. Define $P_x: \partial B \rightarrow \partial T$ by defining $P_x(y)$ to be the intersection of the ray from $x$ to $y$ with $\partial T$. Since $T$ is convex, this is well defined; suppose, by way of contradiction, that some ray emanating from $x$ hits a pair of distinct points $b_1, b_2 \in \partial T$, in this order. Since $x$ is an interior point of $T$, there is some open set $O$ that includes $x$ entirely inside $T$. Now, define the set $$\{t : t = \lambda y + (1- \lambda) b_2, y\in O, \lambda \in (0,1)\} \, .$$ This set is open, and contains $b_1$. By convexity of $T$, it is completely inside $T$, and thus $b_1$ is not a boundary point, contradicting the assumption. This also implies that the map $P_x$ is bijective.
We now use the map $P_x$ to control the $(d-1)$ dimensional Hausdorff measure of the boundary of $T$. To do so, we consider $\{U_i\} \in U_\delta(\partial
B)$ - i.e. a particular covering of the boundary of the Euclidean ball $B$. Without loss of generality, we may assume that all the $U_i$’s are subsets of $\partial B$; otherwise, we simply intersect the $U_i$’s with $\partial B$, possibly decreasing the diameter. Define $\{U_i'\}$ to be the image of the $U_i$’s under $P_x$. The $\{U_i'\}$’s cover $\partial T$, since $P_x$ is bijective, and the diameter of $U_i'$ is no greater than the diameter of $U_i$. Taking a minimizing sequence of coverings of $\partial B$, we conclude that $$H^{d-1}(\partial T) \leq H^{d-1}(\partial B) \, .$$ Note that, formally, it is possible that either of these measures is infinite. However, the Hausdorff measure of the Euclidean ball is well known to be finite [@Federer pg. 171], meaning $H^{d-1}(\partial T)$ is finite.
We now apply these bounds to $S$, the convex subset of a ball of radius $r$. By equivalence of norms, $S$ is also a subset of a Euclidean ball of radius $Cr$ centered at some point in its interior, for some $C$; in fact, for all sufficiently small $z$, both $S_z$ and $S^{(z)}$ are also subsets of this ball. Therefore, using the above inequality, $$H^{d-1}(\{y : \|\partial S - y\| = z\}) = H^{d-1}\left(\partial S_z\right) + H^{d-1}\left(\partial S^{(z)}\right) \leq C H^{d-1}(\partial B)\, .$$ Since B is a ball of radius $Cr$, there exists some constant that depends only on the norm and the dimension such that $$H^{d-1}(\partial B) = C r^{d-1} \, ,$$ for some possibly different $C$. Plugging in these estimates into the Euclidean coarea formula, we get $$\lambda((\partial S)_{\rho/m}) \leq C' \displaystyle \int_{0}^{\rho /m} r^{d-1} dz \leq \frac{C''}{m^d} \, ,$$ where $C''$ and $C''$ are possibly different constants depending only on the dimension and norm. Since $m > (s/r)-1$ and $\tau = \nu (r/2)^d$, we conclude that, for sufficiently large $s$ depending only on $d$, the norm, and $\varepsilon$, $$z\left((\partial S)_{\rho/m}\right) \leq \frac{C''}{\tau m^d} \leq \frac{C'''}{(s-r)^d} \leq \frac{\varepsilon^2}{64}\, ,$$ as required.
As a final preliminary, we prove a lemma about the geometry of maximal clique sets:
\[GeoLemma\] Let $\mathfrak{P}$ be a maximal clique set. Then, for any $\varepsilon >0$, $\mathfrak{U}(\mathfrak{P})$ contains a ball of diameter $(1-\varepsilon)r$, for some sufficiently large $s$.
To prove this lemma, we go through the abstract framework of Hausdorff convergence of subsets of a metric space. Consider an abstract metric space $X$ imbued with metric $\iota$, and, for any $S \subset X$, define the $l$-fattening of $S$ as before, using the metric $\iota$ to measure distance. For any two $A,B \subset X$, the Hausdorff distance is defined as $$\iota_{H} (A,B) := \inf \{ l : A \subset (B)_l,\, B \subset (A)_l\}\, .$$ If $X$ is compact in the topology defined by $\iota$, the space of closed subsets of $X$ makes a compact space with respect to this metric [@Petersen page 294].
Let $A$ be a cube of side $3r$ in the $d$-dimensional torus. Note that $\mathfrak{U}(\mathfrak{R}(A))$ includes a cube of side $(3-2/s)r$. If $s >2$, such a set will also include a ball of diameter $r$, and therefore at least one maximal clique set. For each $s > 2$, let $\mathfrak{P}_s$ be some maximal clique set that is a subset of $\mathfrak{R}(A)$.
Let $W_s = \mathfrak{U}(\mathfrak{P}_s)$. The diameter of $W_s$ is bounded above by $r (1 + C/s)$, and its measure is bounded below by $\tau$ from Lemma \[lmm3\]. Let $\tilde{A}$ and $\tilde{W}_s$ be the same sets, with distances scaled by $1/r$. Then $\tilde{W}_s$ are subsets of a cube of side-length $\le 3$. Their diameter cannot exceed $(1 + C/s)$ and their Lebesgue measure is at least $\nu/2^d $, where the final bound follows from the definition of $\tau$. Note that, after rescaling, every $\tilde{W}_s$ is a subset of the [*same*]{} metric space $\tilde{A}$, which is independent of $n$. Thus, any convergence in the Hausdorff distance associated with $\tilde{A}$ is automatically uniform in $n$. Since $\tilde{A}$ is a compact metric space and $\tilde{W}_s$ is a sequence of closed sets, the compactness result above guarantees that some subsequence $\tilde{W}_{s_k}$ converges to a limit $\tilde{W}$ in the Hausdorff metric. Passing through to the limit, we see that $\tilde{W}$ must have measure at least $\nu/2^d$, and its diameter cannot exceed 1. However, it is also a subset of a cube of side-length $ \le 3$, and therefore we can embed $\tilde{W}$ into the finite dimensional normed space $(\mathbb{R}^d, \| \cdot \|)$ isometrically. Quoting the isodiametric inequality again [@Burago pg 94], $$\lambda(\tilde{W}) \leq \nu \left(\frac{\text{diam} \tilde{W}}{2} \right)^{d}\, .$$ Thus, $\tilde{W}$ satisfies equality for the isodiametric inequality. Luckily, the isodiametric inequality also characterizes all sets that achieve equality as balls in the given norm.
Fix $\varepsilon >0$. For all sufficiently large $k$, the Hausdorff convergence of $\tilde{W}_{s_k}$ guarantees that, for some $B(x)$, a ball of diameter $1$ centered at $x$, we have $$B(x) \subset (\tilde{W}_{s_k})_{\varepsilon/2}\, .$$ Now, consider the set $B(x) \backslash \tilde{W}_{s_k}$. The distance between $x$ and this set must exceed $1/2 - \varepsilon/2$; otherwise, $B(x)$ would not be inside the $\varepsilon/2$-fattening of $\tilde{W}_{s_k}$. Thus, a ball of diameter $1 - \varepsilon$ must be inside $\tilde{W}_{s_k}$. Scaling by $r$ completes the proof of Lemma \[GeoLemma\].
Fix $ \delta > 0 $ and $\varepsilon >0$, and assume that $|E| > (1+ \delta) \mu$. Let $\tilde{\delta} = \delta(1 - \varepsilon/16)$. By \[DeltaCor\], $|E_s| > (1 + \tilde{\delta}) \tilde{\mu}_s$ occurs for some sufficiently large $s$. Let $\tilde{\varepsilon} = \varepsilon/8$. Assume that the event $F$ described in Theorem \[thm2\] happens. For ease of notation, let $q = (2 \tilde{\delta} \tilde{\mu}_s)^{1/2}$. Then there is a maximal clique set $\mathfrak{P}\subseteq T$ such that for all $I\in \mathfrak{P}$, $$\biggl|\frac{\tilde{\tau}_s X_I}{q} - 1 \biggr| < \tilde{\varepsilon}\, .$$ Now, let $A$ be some ball of diameter $(1 - \varepsilon/128) r$ contained in $\mathfrak{U}(\mathfrak{P})$ - for sufficiently large $s$, such as set exists by Lemma \[GeoLemma\]. Pick $S$ to be a convex subset of $A$ with $z(S)\geq \varepsilon/16$. We now wish to show that $$(z(S)- 15 \varepsilon/16) \sqrt{2 \delta \mu} \leq |\chi(S)| \leq (z(S)+ 15\varepsilon/16) \sqrt{2 \delta \mu}\, .$$ This would be sufficient to prove the first stipulation of Theorem \[thm1\] if $A$ were a ball of radius $r$; because of the error in the radius, we must get slightly better bounds and strengthen them later. To get an upper bound, we look at the number of vertices in the outer hull of $S$; the corresponding lower bound will arrive via a bound on the inner hull.
First, we want to get an upper bound on $|\mathfrak{O}(S)|$. By the upper bound on the Lebesgue measure of $\mathfrak{U}(\mathfrak{O}(S))$ from \[FatteningProp\], we conclude that $$\begin{aligned}
|\mathfrak{O}(S)| &\leq (m^d \tau) \left(z(S) + \frac{\varepsilon^2}{64} \right) \\
&\leq z(S) \tilde{\tau}_s \left(1 + \frac{\varepsilon^2}{64 z(S)} \right) \\
&\leq z(S) \tilde{\tau}_s \left(1 + \frac{\varepsilon}{4} \right)\, ,\end{aligned}$$ where the penultimate inequality follows from Lemma \[lmm3\], and the final one from the assumed lower bound on $z(S)$. If we allow every $X_I$ in $\mathfrak{O}(S)$ to take on the maximal value allowed by Theorem \[thm2\], we deduce the inequality $$\begin{aligned}
|\chi(S)| &\leq \left|\chi\left(\mathfrak{U}(\mathfrak{O}(S))\right)\right|\\
&\leq z(S) \tilde{\tau}_s \left(1 + \frac{\varepsilon}{4} \right) \left[\left(\frac{q}{\tilde{\tau}_s}\right)\left(1 + \tilde{\varepsilon}\right)\right]
\\
&\leq z(S) q \left(1 + \frac{\varepsilon}{2}\right)\, .\end{aligned}$$ This is nearly the desired upper bound; we simply need to replace $q$ with $\sqrt{2 \delta \mu}$. By appealing to \[lmm3\], we see that $$q = (2 \tilde{\delta} \tilde{\mu}_s)^{1/2} \leq \sqrt{2 \delta \mu} \left(1 + \frac{C}{s}\right)^{1/2}\, .$$ Increasing $s$ sufficiently to ensure that $$\left(1 + \frac{C}{s}\right)^{1/2} \leq \left(1 + \frac{\varepsilon}{4}\right)$$ and substituting into the earlier inequality, we produce the desired upper bound on $|\chi(S)|$.
The lower bound follows similarly. First, we get a lower bound on $|\mathfrak{R}(S)|$: $$|\mathfrak{R}(S)| \geq \left(z(S) - \frac{\varepsilon^2}{64}\right)(m^d \tau) \geq z(S) \tilde{\tau}_s \left(1 + \frac{C}{s}\right)^{-1} \left(1 -
\frac{\varepsilon}{4}\right) \,$$ where we appeal to \[FatteningProp\] for the lower bound on the Lebesgue measure of $\mathfrak{U}(\mathfrak{R}(S))$. Since $S\subset A$, $\mathfrak{R}(S) \subset \mathfrak{P}$. If every $X_I$ is bounded below by the minimal value in $\mathfrak{P}$, we get the following lower bound for $|\chi(S)|$: $$|\chi(S)| \geq \left|\chi\left(\mathfrak{U}[\mathfrak{R}(S)]\right)\right| \geq z(S) \tilde{\tau}_s \left(1 + \frac{C}{s}\right)^{-1} \left(1 -
\frac{\varepsilon}{4}\right) \left[\left(\frac{q}{\tilde{\tau}_s}\right)\left(1 - \tilde{\varepsilon}\right)\right]\, .$$ The lower bound on $q$ $$q = (2 \tilde{\delta} \tilde{\mu}_s)^{1/2} \geq \biggl(\frac{2 \delta \mu}{1 - \varepsilon/16}\biggr)^{1/2}$$ follows from the Lemma \[lmm3\] bound $\tilde{\mu}_s \geq \mu$ and the definition of $\tilde{\delta}$. Substituting this in and increasing $s$ sufficiently gives the desired lower bound.
As mentioned above, this nearly completes the first stipulation of Theorem \[thm1\]; the only difference is that the ball $A$ is of diameter $(1 -
\varepsilon/128)r$ instead of $r$. Let $A'$ be an arbitrary ball of diameter $r$ containing $A$, and pick a convex set $S'$ in $A$. By definition, $z(A'/A) \leq \varepsilon/32$. Defining $S = S' \cap A$, all the previous assertions follow. Furthermore, the lower bound $$|\chi(S')| \geq |\chi(S)| \geq (z(S)- 15\varepsilon/6) \sqrt{2 \delta \mu}$$ follows from before. Finally, $z(S) \geq z(S') - \varepsilon/32$, as $z(S'/S) \leq z(A'/A)$, which implies the lower bound $$|\chi(S')| \geq (z(S')- \varepsilon) \sqrt{2 \delta \mu}$$ For the upper bound, we note that $$\begin{aligned}
|\chi(S')| &\leq |\chi(S)| + |\chi(\mathfrak{O}(S'/S))| \\
&\leq \left(z(S) + 15 \varepsilon/16 \right) \left(\sqrt{2 \delta \mu}\right) + |\chi(\mathfrak{O}(S'/S))|\, .\end{aligned}$$ Again, assuming all elements of $\mathfrak{O}(S'/S)$ intersect the indices with largest density, we can conclude that $$|\chi(\mathfrak{O}(S'/S))| \leq z(S'/S)\tilde{\tau}_s\left(1 + \frac{\varepsilon}{4}\right)\left(\frac{q}{\tilde{\tau}_s}\right)\left(1 +
\tilde{\varepsilon}\right) \leq \left(\frac{\varepsilon}{16}\right) (\sqrt{2 \delta \mu})$$ following the same derivations as above. Therefore, $$(z(S')- \varepsilon) \sqrt{2 \delta \mu} \leq |\chi(S')| \leq (z(S')+ \varepsilon) \sqrt{2 \delta \mu}$$ for [*any*]{} convex set $S'$ of $A'$ (including $A'$ itself), proving the first stipulation of \[thm1\].
For the second part of Theorem \[thm1\], consider a convex set $S \subset A^c$, with $z(S) > \varepsilon/16$ that lies completely inside a ball of radius $r$. For sufficiently large $s$, $\mathfrak{R}(S)$ will be completely disjoint from $\mathfrak{P}$; however, $\mathfrak{O}(S)$ may not be. Thanks to the bounds on the measure of the outer and inner hulls, $$|\mathfrak{O}(S)/\mathfrak{R}(S)| \leq \left(\frac{\varepsilon^2}{32} \right) m^d \tau \leq \left(\frac{\varepsilon^2}{26} \right) \tilde{\tau}_s\, .$$ To get an upper bound on $|\chi(S)|$, we assume that $\mathfrak{O}(S)$ has the maximal intersection with $\mathfrak{P}$, and that the remaining elements take on the maximal value allowed by Theorem \[thm2\]. Specifically, this gives $$\begin{aligned}
|\chi(S)| &\leq |\chi(\mathfrak{U}(\mathfrak{O}(S)))| \\
&\leq |\mathfrak{O}(S) \cap\mathfrak{P}| \left(\frac{q}{\tilde{\tau}_s}\right)\left(1 + \tilde{\varepsilon}\right) + |\mathfrak{O}(S)|
\left(\frac{\tilde{\varepsilon} q}{\tilde{\tau}_s}\right) \\
&\leq z(S) \left[\frac{\varepsilon^2}{26 z(S)}(1 + \tilde{\varepsilon}) + K\left(1 + \frac{\varepsilon}{4}\right) 2 \tilde{\varepsilon}\right] q\end{aligned}$$ using the earlier bound on the cardinality of the outer hull and the upper bound on the $X_J$, $J \not \in \mathfrak{P}$ from Theorem \[thm2\]. Thanks to the lower bound on $z(S)$ and the definition of $\tilde{\varepsilon}$, the bracketed expression can be bounded above by $3 \varepsilon/4$. Using the bound $q < (1 + \varepsilon/4) \sqrt{2 \delta \mu}$ for all large $s$, derived above, and dividing through by $\sqrt{2\delta \mu}$, gives the desired result. This completes the proof of Theorem \[thm1\].
Proof of Theorem \[thm2\]
=========================
The bulk of the remaining length of the paper is dedicated to the proof of Theorem \[thm2\]. The proof is organized as follows: First, we use concentration inequalities and large deviations estimate to show that, in the $s$-graded model conditioned to have too many edges, three inequalities hold with high probability. Next, we show that these inequalities imply the existence of a small set of indices such that almost all the “extra” edges have both endpoints in the union of its associated $A_I$’s, and satisfies a convex upper bound. The next two subsections require delicate estimates which produce a subset which satisfies a similar quadratic lower bound, a termwise lower bound, and an additional upper bound on the total number of vertices in the set. Finally, simple convex analysis will show that this set will be a maximal clique set, and that the vertices are roughly equidistributed among the $A_I$’s in this set.
Probabilistic Analysis {#probanalysis}
----------------------
For this section, $C$ will indicate any multiplicative constant that is uniformly bounded in $n$. It may depend on all other parameters, including $s$, and may take on different values from line to line.
In this section, we will estimate the probability of four rare events. First, we will produce a lower bound for the probability of the conditioning event – namely, that the number of edges exceeds its mean by a multiplicative factor of $(1 + \tilde{\delta})$. Next, we prove upper bounds for three events that will constrain the large deviation event. One will involve the number of $I$’s for which $X_I$ is big, the second will bound the distribution of vertices among the large $X_I$’s, and the third will control the large deviation probability of the edge count between the small $X_I$’s. We then combine all the estimates in the proof of \[HighProb\], the essential probabilistic proposition of this paper.
Let $L$ be the event $\{ |E_s| > (1+ \tilde{\delta}) \tilde{\mu}_s\}$. As before, let $q = (2 \tilde{\delta} \tilde{\mu}_s)^{1/2}$ and $w = \tilde{\tau}_s
\mathcal{D}$. Our first goal is to produce a lower bound on the probability of $L$:
\[LLowerBound\] If $L$, $q$ and $w$ are defined as above, then there exists a $C$ such that $$\mathbb{P}(L) \geq \exp\left( - q (\log (q/w) - 1) - C n^z \log n\right)\, ,$$ where $z= \max\{p/4, 3p/4-1/2\}$.
Pick a maximal clique set $\mathfrak{K}$ and define the event $H$ by $$H := \left\{\forall I \in \mathfrak{K}, X_I \geq \left\lceil\frac{q + n^z}{\tilde{\tau}_s} \right\rceil \right\} \, .$$ Note that the choice of $z$ is somewhat arbitrary, and is motivated below; for now, it is important that, since $\log \tilde{\mu}_s \sim p\log n$ as $n \to \infty$ (by the definition of $p$), implies that the correction term $n^z$ is of lower order than $q$, because $\max\{p/4,
3p/4-1/2\} < p/2$ for all admissible values of $p$.
Assume that $H$ occurs. This implies that the number of edges (in the graph $G_s$) with both endpoints in $\mathfrak{U}(\mathfrak{K})$ is at least $${q + n^z \choose 2} \geq \delta \tilde{\mu}_s + \frac{q n^z}{2}$$ Therefore, if $H_{\mathfrak{K}}$ occurs then $L$ would be true if the total number of edges with at most one endpoint in $\mathfrak{U}(\mathfrak{K})$ would exceed $\tilde{\mu}_s -(qn^z)/2$. Let $|E_s|'$ be the number of edges with neither endpoints in $\mathfrak{U}(\mathfrak{K})$. Then $$\mathbb{E}(|E_s|') \geq \displaystyle \sum_{I \, : \, d(I,\mathfrak{K}) > s} \mathbb{E} \biggl[{X_I \choose 2} + \frac{1}{2} \sum_{J\, : \, 0 < d(I,J)
\leq s} X_I X_J \biggr]\, .$$ This is a lower bound as we ignore pairs of nonzero elements which are close to $\mathfrak{K}$ but not in $\mathfrak{K}$. The set $\{ I :
d(I,\mathfrak{K}) \leq s\}$ has diameter $\le 3s$, and therefore its cardinality must be some constant independent of $n$. Therefore by , $$\mathbb{E}(|E_s|') \geq \left(m^d - C \right) \frac{|N_I| n^2 }{2 m^{2d}} = \tilde{\mu}_s - \frac{C \tilde{\mu}_s}{m^d}\, .$$ We now wish to bound the probability $$\mathbb{P}(|E_s|' < \tilde{\mu}_s - q n^z /2) \leq \mathbb{P}\left(|E_s|' < \mathbb{E}(|E_s|') - q n^z /2 + \frac{C \tilde{\mu}_s}{m^d} \right)\, ,$$ where the inequality follows from the above relation between the mean of $|E_s|'$ and $\tilde{\mu}_s$. By Lemma \[lmm3\], $\tilde{\mu}_s/m^d$ grows as $n^{2(p-1)}$ (on the logarithmic scale). If $p<1$, this is a vanishing quantity, and therefore $$q n^z /2 - \frac{C \tilde{\mu}_s}{m^d} \leq q n^z/4\, .$$ If $p\geq 1$, then $qn^z$ grows as $n^{5p/4 - 1/2}$. Since $p < 2 -\delta^*$ by the restrictions on $r$, $2(p-1) < 5p/4 -1/2$, and, for sufficiently large $n$, the inequality above still holds.
Using the sum definition of $|E_s|$, a straightforward calculation can show that $$\text{Var}(|E_s|') \leq \text{Var}(|E_s|) \leq C m^d \left(\mathcal{D}^3 + \mathcal{D}^2\right)\, .$$ Referring back to the growth bounds on $m^d$ and $\mathcal{D}$, we see that the variance grows as $n^p$ if $p<1$ and $n^{2p-1}$ if $p \geq 1$. By Chebyshev’s inequality, $$\mathbb{P}(|E_s|' < \mathbb{E}(|E_s|') - q n^z /4) \leq \frac{16 \text{Var}(|E_s|')}{(q n^z)^2}\, .$$ For all value of $p$ of interest, this fraction vanishes as $n^{-p/2}$, up to possible logarithmic factors. Thus, the probability of the event $$\mathbb{P}(|E_s|' \geq \tilde{\mu}_s - q n^z /2) > 1 - \varepsilon$$ for any $\varepsilon$ positive, for sufficiently large $n$.
Since the above event implies $L$ once $H$ occurs, this allows us to conclude that $$\mathbb{P}(L) \ge \mathbb{P}(H) \mathbb{P}(L\mid H) \geq (1 - \varepsilon) \, \mathbb{P}(H)\, .$$ Thus, the lower bound on the probability of $L$ will come from a good lower bound on the probability of $H$. By definition, $$\begin{aligned}
\mathbb{P}(H) &= \left[\mathbb{P}\left(X_I \geq \left\lceil\frac{q + n^z}{\tilde{\tau}_s} \right\rceil\right)\right]^{\tilde{\tau}_s} \\
& \geq \left[\mathbb{P}\left(X_I = \left\lceil\frac{q + n^z}{\tilde{\tau}_s} \right\rceil\right)\right]^{\tilde{\tau}_s} \\
& \geq \left( e^{-\mathcal{D}} \mathcal{D}^{(q+ 2n^z)/\tilde{\tau}_s }/\left[(q +2n^z)/\tilde{\tau}_s \right]!\right)^{\tilde{\tau}_s} \,\end{aligned}$$ where the final inequality follows by removing the ceiling function and compensating by adding a factor of two to the correction term. Now, by Stirling’s approximation, $$\left(\left[(q +2n^z)/\tilde{\tau}_s \right]!\right)^{-1} \geq \exp\left(-\left(\frac{q +3n^z}{\tilde{\tau}_s}\right) \left[\log \left(\frac{q
+3n^z}{\tilde{\tau}_s}\right) - 1\right] \right) \, ,$$ where the changed constant in front of the $n^z$ term is to compensate for the omitted polynomial term in the approximation. Substituting this above gives $$\mathbb{P}(H) \geq \exp \left(- \tilde{\tau}_s \mathcal{D} - (q + 3n^z) \left[\log \left(\frac{q + 3n^z}{\tilde{\tau}_s \mathcal{D}}\right)
-1\right]\right)\,.$$ Since $n^z$ and $\tilde{\tau}_s \mathcal{D}$ are both lower order correction, we can find an absolute constant $C$ such that the desired bound holds, completing the proof.
Let $a = \delta^*/25$ and $M = \max\{ \mathcal{D} n^a, n^a \}$ We will say an index $I \in T$ is in the [*bulk*]{} if $$X_I \leq M\, .$$ First, we wish to show that not too many indices are outside the bulk.
\[BulkCardinality\] Let $$\alpha = \min\{1 - p/2 - a/2,\, p/2 - a/2\}$$ and define $A$ to be the event $$\{\exists S \subset T, |S| > n^\alpha \text{ such that } \forall I \in S,\, X_I > M\}\, .$$ Then, for some $C >0$, $$\mathbb{P}(A) \leq \exp(-Cn^{p/2+a/2})$$
Note that it is sufficient to show that no such set of cardinality exactly $\lceil n^\alpha \rceil$ exists. We can bound the probability of this event in a straightforward manner. First, we establish the classical large deviation bound for Poisson random variables. To do so, note that $$\mathbb{E}[\exp(\lambda X_I)] = e^{-\mathcal{D}} \displaystyle \sum_{k=0}^\infty \frac{\exp(\lambda k) \mathcal{D}^k}{k!} = \exp\left[ \mathcal{D}
\left(e^{\lambda} -1\right) \right]$$ By Chebyshev’s inequality, this implies that, for any positive $\lambda$, $$\mathbb{P}(X_I > t) \leq \exp \left[\mathcal{D} \left(e^{\lambda} -1\right) - \lambda t \right]$$ Setting $\lambda = \log(t/\mathcal{D})$ gives $$\label{ChernoffUp}
\mathbb{P}(X_I > t) \leq \exp (- t [\log (t/\mathcal{D}) -1] - \mathcal{D}), \, t > \mathcal{D} \, .$$ A similar procedure, using function $\mathbb{E}[e^{-\lambda X_I}]$, leads to $$\label{ChernoffDown}
\mathbb{P}(X_I < t) \leq \exp (- t [\log (t/\mathcal{D}) -1] - \mathcal{D}), \, t < \mathcal{D} \, .$$ We now return to bounding the probability of $A$. If $p \geq 1$, we get that $$\label{lower1}
\mathbb{P}(A) \le {m^d \choose n^\alpha} \mathbb{P}(X_I > M)^{n^\alpha} \leq m^{d n^\alpha} \exp(-n^{\alpha} M \log (M/\mathcal{D}) )\, ,$$ The polynomial rate of growth of $n^{\alpha + a}\mathcal{D}$ is $p/2 + a/2$, whereas $n^\alpha$ grows as $1 -p/2 - a/2$, which is smaller for all $p \geq
1$. Therefore, replacing $M \log (M/\mathcal{D})$ by $n^a \mathcal{D}$, which is strictly smaller, implies that $$\mathbb{P}(A) \leq \exp(-n^{\alpha + a} \mathcal{D} + d n^\alpha \log m ) \leq \exp(-Cn^{p/2+a/2})$$ If $p< 1$, the same calculation gives $$\label{lower2}
\mathbb{P}(A) \leq {m^d \choose n^{p/2 -a/2}} \exp(-n^{p/2 + a/2}) \leq \exp(-Cn^{p/2 +a/2})\, .$$
Next, we want some control over the behavior of indices outside the bulk. Define $$Y_I = X_I \big(\log (X_I/\mathcal{D}) -1\big) + \mathcal{D} \, ,$$ with the convention that $0 \cdot \log 0 = 0$. Note that $Y_I = \mathcal{I}(X_I)$, where $\mathcal{I}$ is the rate function of a Poisson random variable of mean $\mathcal{D}$. Therefore, $\mathbb{P}[Y_I > t]$ should vanish as $\exp(-t)$, by “inverting" the rate function. We formalize this notion in the lemma below:
\[ExpMomentGeneratingY\] For any $\mathcal{D}$, any positive $\lambda <1$, and some $C$ uniform in $n$, $$\mathbb{E}\left[\exp(\lambda Y_I)\right] \leq \frac{C}{1 - \exp( C(\lambda -1)/ \log \mathcal{D})}$$
First note that $Y_I \geq 0$, as the function $f(x) = x [\log (x/\mathcal{D}) - 1] + \mathcal{D}$ achieves its minimum at $x = \mathcal{D}$. This means that, in general, $f(x)$ is not invertible for all positive integer values of $x$. Therefore, we define two inverses. First, let $$g_1(x): [0,\mathcal{D}] \rightarrow [0,\mathcal{D}] \text{ be a function such that } (f \circ g_1) (x) = x\, .$$ Note that this function is decreasing, with $g_1(0) = \mathcal{D}$ and $g_1(\mathcal{D}) = 0$. For any $x > \mathcal{D}$, we say that $g_1(x) = -\infty$. We define $g_2$, the second inverse, similarly, except its range is defined to be $(\mathcal{D},\infty)$. This inverse is strictly increasing. We use and to deduce $$\mathbb{P}(Y_I > t) = \mathbb{P}(X_I < g_1 (t)) + \mathbb{P}(X_I > g_2 (t)) \leq 2 e^{-t}\, .$$ Note that, if $t > \mathcal{D}$, the first term contributes nothing to the probability.
These estimates allow us to bound the moment generating function of $Y_I$: let $\mathcal{A}$ be the set of atoms of $Y_I$. Then, for any $\lambda <1$, $$\mathbb{E}\left[\exp(\lambda Y_I)\right] \leq \displaystyle \sum_{i \in \mathcal{A}} 2e^{-i+ \lambda i}\, .$$ If $p < 1$, $\mathcal{A}$ is more sparse than the integers, in the sense that $|\mathcal{A} \cap [0,n] | \leq n$. Thus, we can increase the expectation by taking the sum over the integers, and possibly adding a multiplicative constant. If $\mathcal{D} \rightarrow \infty$ (which would require $p \geq 1$), the minimal distance between adjacent atoms in $\mathcal{A}$ is $C/\log \mathcal{D}$. Using these results, we conclude that, if $p <1$, $$\label{ExpSmall}
\mathbb{E}\left[\exp(\lambda Y_I)\right] \leq C \displaystyle \sum_{i=1}^\infty e^{(\lambda-1) i} = \frac{C}{1 - \exp(\lambda -1)} \, .$$ Meanwhile, when $p \geq 1$, can only produce the inferior bound $$\label{ExpLarge}
\mathbb{E}\left[\exp(\lambda Y_I)\right] \leq C \sum_{i =1}^\infty e^{C (\lambda-1) i/\log \mathcal{D}} \leq \frac{C}{1 - \exp( C(\lambda -1)/ \log
\mathcal{D})} \, .$$ The second bound, which holds even if $p<1$, is the one we required in the lemma.
The sum of the $Y_I$’s over indices [*not*]{} in the bulk will play a critical role in later analysis. We next produce an upper bound on the probability that there exists a small set of indices for which the sum of $Y_I$’s is particularly large. Intuitively, \[BulkCardinality\] shows that the complement of the bulk is not particularly large, so this estimate will control the desired quantity. Furthermore, the exponential tail of $Y_I$ allows us to produce a very good upper bound on that sum.
\[YIBound\] Let $Y_I$ and $\alpha$ as above, and define $\beta := p/2 - a/4$. Define the event $$B := \biggl\{ \exists S \subset T, |S| \leq n^{\alpha} \text{ such that } \displaystyle \sum_{I \in S} Y_I > t \biggr\}\, ,$$ where $$t = q (\log(q/w) -1) + n^\beta \, .$$ Then $$\mathbb{P}(B) \leq \exp( -t + n^\beta/2) \, .$$
We now use a Chernoff-like strategy to bound the probability of the event $$\left\{ \displaystyle \sum_{I \in S} Y_I > t \right\}$$ for some $S \subset T$ of cardinality at most $n^{\alpha}$. By Chebyshev’s inequality $$\begin{aligned}
\mathbb{P}\Big(\displaystyle \sum_{I \in S} Y_I > t\Big) & \leq \frac{\big(\exp(\lambda Y_I)\big)^{n^\alpha}}{\exp(\lambda t)} \\ &\leq \left(\frac{C}{1 -
\exp(C (\lambda -1)/ \log \mathcal{D}) } \right)^{n^\alpha} e^{-\lambda t} \, .\end{aligned}$$ Note that we used the estimate in \[ExpMomentGeneratingY\] in that calculation.
We now set $\lambda = 1 - n^{\alpha} /{t}$. This leaves an exponential in $n^\alpha/(t \log \mathcal{D})$, which is vanishing in $n$. Therefore, we can bound $$\begin{aligned}
\left(1 - \exp(-C n^{\alpha}/(t \log \mathcal{D}))\right)^{-1} & \leq \left(C n^{\alpha}/(t \log \mathcal{D}) - (C n^{\alpha}/(t \log \mathcal{D}) )^2\right)^{-1}\\ &
\leq (C t \log \mathcal{D}) /n^\alpha + 2\end{aligned}$$ for all sufficiently large $n$. Therefore, $$\mathbb{P}\Big(\displaystyle \sum_{I \in S} Y_I > t\Big) \leq \left( C (t \log \mathcal{D}) / n^{\alpha} + 2\right)^{n^{\alpha}} \cdot e^{-t + n^{\alpha}} \leq \exp\left(- t + C n^{\alpha} \log n\right)\, .$$
To calculate the probability of the event $B$ defined in the lemma, we invoke the union bound: $$\mathbb{P}(B) \leq {m^d \choose n^{\alpha}} \exp\left( -t + C n^{\alpha} \log n\right) \, .$$ We bound the combinatorial factor $${m^d \choose n^{\alpha}} \leq m^{d n^{\alpha}} \leq \exp^{ C n^{\alpha} \log n} \, ,$$ for some $C$ independent of $n$, since $m^d$ is a polynomial in $n$. Since $\beta > \alpha$, we can be sure that $C n^{\alpha} \log n < n^\beta /2 $ for all sufficiently large values of $n$. Therefore, $$\mathbb{P}(B) \leq \exp ( - t + n^\beta/2)$$ as required.
The final probabilistic step is to show that, conditional on $L$, the bulk does not contain very many extra edges. Let $$\hat{X}_I := X_I \cdot 1_{X_I \leq M}$$ and $\hat{|E_s|}$ be the define analogously with $|E_s|$ by replacing $X_I$ with its truncated version (Recall that $M = \max\{n^a, \mathcal{D} n^a\}$). In other words, $\hat{|E_s|}$ is the version of $G_s$ obtained after deleting all vertices in boxes that satisfy $X_I > M$. Fix $\gamma = p - 2a$, and consider the event $$D = \big\{ |\hat{E}_s| - \tilde{\mu}_s > n^\gamma\big\}\, .$$ The argument for controlling the probability of $D$ splits into two regimes. We first treat the situation in which $\mathcal{D}$ is vanishing as $n$ grows; this covers the case $p<1$ and a subset of the possible values of $p=1$. We wish to prove the following lemma:
\[ConcIneqSmall\] Let $D$ be as above, and assume that $\mathcal{D}$ is vanishing as $n$ grows. Then, for some $C>0$, $$\mathbb{P}(D) \leq \exp\left( - C n^{2 \gamma - p - 8a}\right)$$
With our additional assumption of $\mathcal{D}$, we know that $M=n^a$. Therefore, $\hat{|E_s|}$ has a Lipschitz constant of $C n^{2a}$ in all its coordinates with respect to the Hamming metric when seen as a function of the $\hat{X}_I$’s.
We use Talagrand’s convex concentration inequality [@Talagrand2 Theorem 4.1.1]. First, let us define the setting: let $\Omega = \prod_{i=1}^N
\Omega_i$, where $\Omega_i$ are all probability spaces, the measure on $\Omega$ is the product measure, and $X$ is a random variable. For a set $A \subset
\Omega$, define the set $$U_A(x):= \{ \{s_i\} \in \{0,1\}^N : \exists y \in A, s_i = 0 \implies x_i = y_i\}\, .$$ Let $V_A(x)$ be the convex hull of $U_A(x)$, and $d_c(A,x)$ is the $\ell^2$ distance of $V_A(x)$ to the origin. For any set $A$, we denote $A_t$ be the $t$ blowup of $A$ with respect to this metric, i.e. $$A_t := \{ x \in \Omega: d_c(A,x) \leq t\}\,.$$ We can now state the inequality:
If $\Omega$, $\mathbb{P}[\cdot]$, $A$ and $A_t$ are as above, then $$\mathbb{P}[A] \left(1 - \mathbb{P}[A_t]\right) \leq e^{-t^2/4} \, .$$
We will not apply this theorem directly; instead, we use a corollary of this theorem frequently used in discrete settings [@Spencer Theorem 7.7.1]. To do so, we consider $f$, a function from the natural numbers to the natural numbers. We say that $f$ is a witness function for $X$ if, whenever $X(\omega) \geq t$, there exists $I \subset [n]$ with $|I| \leq f(t)$, such that every $\omega'$ that agrees with $\omega$ in all $i \in I$ has $X(\omega') \geq t$. Furthermore, we assume that $X(\omega)$ is $K$-Lipschitz with respect to the Hamming distance — that is, $|X(\omega) - X(\omega')|
\leq K$ whenever $\omega$ and $\omega'$ differ in at most one coordinate.
Let $\Omega$ be a product space, and $X$ a real valued function on $\Omega$ with Lipschitz constant $K$ with respect to the Hamming distance. If $f$ is witness function for $X$ as above, then, for any $b$ and $t$, $$\mathbb{P}[X > b + t K\sqrt{f(b)}] \, \mathbb{P}[X \le b ] \leq \exp (-t^2/4)\, .$$
We now apply this theorem on $X=\hat{|E_s|}$. Since each coordinate is bounded above by $n^a$, $X$ is Lipschitz with $K = C n^{2a}$. The function $f(t)
= 2t$ is a witness function for $\hat{|E_s|}$; to see this, note that $\hat{|E_s|}$ is the edge count of the $s$-graded geometric random graph, after we remove any $X_I$ with very high density. As such, we can “witness” the existence of $t$ edges by finding at most $2t$ vertices; the flexibility of the setup allows us to pick these vertices judiciously, avoiding all the isolated ones. Finding $2t$ vertices will require at most $2t$ distinct coordinates, if each one of them vertices lies in a distinct $A_I$.
We apply the theorem with $b = \tilde{\mu}_s + qn^z$ and $$t = \frac{n^{\gamma} - q n^z}{n^{2a} \sqrt{\tilde{\mu}_s + qn^z}}\, ,$$ with $z$ defined as in the beginning of the section, and deduce that $$\begin{aligned}
\mathbb{P}(D) \mathbb{P}(\hat{|E_s|} \leq \tilde{\mu}_s + q n^z) & \leq \exp\left(-\frac{C[n^{\gamma} - q n^{z}]^2 }{(\tilde{\mu}_s + q n^z)
n^{4a}}\right) \\ & \leq \exp\left(-C' n^{2 \gamma - p -8a} \right)\, ,\end{aligned}$$ where the final inequality follows because $q n^z$ is a lower order correction to $\tilde{\mu}_s$, by definition of $z$.
Note that is inequality is only possible when $f(\tilde{\mu}_s) \leq m^d$; $f(\tilde{\mu}_s)$ represents the number of coordinates necessary to witness the property $\hat{|E_s|}$, so it must be bounded by the total number of random variables available. Since $\mathcal{D}$ is vanishing, this is guaranteed to hold for all sufficiently large values of $n$.
Now, we must bound the probability that $\hat{|E_s|}$ is below $\tilde{\mu}_s + qn^z$. This is not a rare event: the mean of $\hat{|E_s|}$ is strictly smaller than the mean of $|E_s|$, and its variance is essentially the same. To see this second fact, note that $$\text{Var}[\hat{|E_s|}] \leq \mathbb{E}\left[(\hat{|E_s|} - \tilde{\mu}_s)^2 \right] \, ,$$ since the function $t \rightarrow \mathbb{E}[(X-t)^2]$ is minimized at the mean of $X$ for any random variable. On the event $\{X_I = \hat{X}_I , \, \, \forall I \in T\}$, we can repeat the computation used in \[LLowerBound\] to show that the variance is bounded above by a constant multiple of $\text{Var}[|E_s|]$. Meanwhile, thanks to the union bound, the complement occurs with probability of at most $m^d \exp ( -C n^a \log n)$. Since the maximal value of $(\hat{|E_s|} - \tilde{\mu}_s)^2$ is $m^{2d} n^{4a}$ by the coordinate-wise bound, we deduce that $$\text{Var}[\hat{|E_s|}] \leq \text{Var}[|E_s|] + m^{3d} n^{4a} \exp(- Cn^a \log n) \leq C m^d \mathcal{D}^2\, ,$$ where the final inequality holds for all sufficiently large values of $n$, as in the computation in \[LLowerBound\]. Repeating the Chebyshev argument used in that proof, we can deduce that, for any $\varepsilon >0$, there exists a sufficiently large $n$ to ensure $$\mathbb{P}(\hat{|E_s|} \leq \tilde{\mu}_s + q n^z) > 1 - \varepsilon \, .$$ This immediately implies $$\mathbb{P}(D) \leq \exp\left(-C n^{2 \gamma - p -8a} \right) \, ,$$ completing the proof.
For the case $\mathcal{D}$ being uniformly bounded from below, we use a different set of tools. Instead of Talagrand’s inequality, the main tool of this proof is the Azuma-Hoeffding inequality. Unfortunately, if $\mathcal{D}$ grows with $n$, a naive application of the inequality does not yield a sufficiently good bound for our purposes; the Lipshitz constant of $|\hat{E}_s|$, which is at most $C\mathcal{D}^2n^{2a}$, is too large. To get around this, we create a larger probability space, and calculate probabilities on this space. We will only apply the Azuma-Hoeffding inequality after we eliminate some pathological behavior by other means. Formally, we prove the lemma:
\[ConcIneqLarge\] If $\mathcal{D}$ is bounded above and below uniformly in $n$, then there exists a constant $C'$ such that $$\mathbb{P}(D) \leq \exp\left(-\frac{C' n^{2\gamma}}{m^d n^{4a}} \right)\, .$$ If $\mathcal{D}$ grows as $n$ grows, then $$\mathbb{P}(D) \leq \exp\left(-\frac{C' n^{2\gamma}}{m^d \mathcal{D}^3 n^{4a}} \right) \, .$$
We first consider the case in which $\mathcal{D}$ is bounded. Then $|\hat{E}_s|$ has Lipshitz constant of at most $C \mathcal{D}^2 n^{2a}$. By a naive application of Azuma-Hoeffding, we conclude that $$\begin{aligned}
\mathbb{P}(D) & \leq \mathbb{P}\left(\hat{|E_s|} - \mathbb{E}[\hat{|E_s|}] > n^\gamma \right) \\ & \leq \exp\left(-\frac{C n^{2\gamma}}{2m^d
\mathcal{D}^4 n^{4a}}\right) \\
& \leq \exp\left(-\frac{C' n^{2\gamma}}{m^d n^{4a}} \right) \, ,\end{aligned}$$ where the first inequality follows from the fact that $\tilde{\mu}_s > \mathbb{E}[\hat{|E_s|}]$, the second is Azuma-Hoeffding, and the final follows because $\mathcal{D}$ is uniformly bounded in $n$. Note that, if $\mathcal{D}$ was growing with $n$, the probability will vary with the [*fourth*]{} power of $\mathcal{D}$, and not the third power, as required by the lemma.
We now turn to the final case. The measure of each $A_I$ is $1/m^d= \mathcal{D}/n$, which is much greater than $1/n$ by assumption on $\mathcal{D}$. We partition $A_I$ into many sets of measure $1/n$; formally, let $\{F_{I,t}\}$, for natural $t \leq \lceil \mathcal{D} \rceil$ be a set of disjoint subsets of $A_I$ such that $\lambda(F_{I,t}) = 1/n$ for every $t \leq \lceil \mathcal{D} \rceil -1$, and $$\bigcup_{t} F_{I,t} = A_I \, .$$ Note that the measure of the final $F_{I,t}$ will be strictly smaller than $1/n$, unless $\mathcal{D}$ is an integer. We define $W_{I,t} =
|\chi(F_{I,t})|$.
Clearly, $\sum_{t} W_{I,t} = X_I$. Define $\overline{|E_s|}$ as (yet another!) truncation of $|E_s|$ - this time, we truncate each $W_{I,t}$ at $n^a/2$. Formally, let $\overline{W_{I,t}} = W_{I,t} \cdot 1_{W_{I,t} < n^a/2}$, and define $\overline{|E_s|}$ by replacing each $X_I$ in the definition of $|E_s|$ by $\sum_t \overline{W_{I,t}}$. Every configuration of this truncation is also allowed in $\hat{|E_s|}$, but $\overline{|E_s|}$ has a far smaller Lipshitz constant with respect to the $W_{I,t}$’s. Specifically, each $W_{I,t}$ interacts with at most $|N_I| (\mathcal{D} + 1)$ different variables, each one at most $n^{a}/2$. Thus, the Lipshitz constant is at most $C \mathcal{D} \cdot n^{2a}$ in the Hamming distance in the larger space. Of course, we pay two prices for this better truncation: first, the splitting procedure increases the dimensionality of the probability space. More importantly, we must control both $\overline{|E_s|}$ and $\hat{|E_s|} - \overline{|E_s|}$ to get control of the probability of $D$. The first task is a straightforward application of Azuma-Hoeffding: $$\begin{aligned}
\mathbb{P}\left[\overline{|E_s|} - \tilde{\mu}_s > n^{\gamma}/2 \right] & \leq \mathbb{P}\left[\overline{|E_s|} - \mathbb{E}[\overline{|E_s|}] > n^{\gamma}/2 \right] \\ & \leq \exp\left(- \frac{C n^{2\gamma}}{m^d \mathcal{D}^3 n^{4a}}\right) \, ,\end{aligned}$$ using the fact that there are fewer than $m^d (\mathcal{D} +1)$ coordinates, each with Lipshitz constant $C \mathcal{D} \cdot n^{2a}$. This bound is sufficiently good for our purposes, as the denominator varies with third power of $\mathcal{D}$.
By partitioning, $$\mathbb{P}[D] \leq \mathbb{P}\left[\overline{|E_s|} - \tilde{\mu}_s > n^{\gamma}/2 \right] + \mathbb{P}\left[D, \overline{|E_s|} - \tilde{\mu}_s < n^{\gamma}/2 \right] \, .$$ The second event on the righthand side is dominated by the event $G := \{\hat{|E_s|} - \overline{|E_s|} > n^\gamma/2\}$. We now turn to bounding the probability of this event. The difference between the two random variables is given by configurations in which at least $W_{I,t}$ is larger than $n^{a}/2$. In fact, $$\hat{|E_s|} - \overline{|E_s|} < \displaystyle \sum_{(I,t)} \left[ \left( W_{I,t} \cdot 1_{W_{I,t} > n^a/2} \right) \cdot \sum_{J : d(I,J) \leq s} \left( X_J \cdot 1_{X_J < \mathcal{D} n^a} \right) \right]\, .$$ While the random variables in the expression above are far from independent, we can replace the second sum over the $X_J$’s by $|N_I| \mathcal{D} n^a$, the upper bound imposed on it by the indicator random variables involved. Therefore, $$\mathbb{P}[G] \leq \mathbb{P}\left[\displaystyle \sum_{(I,t)} \left(W_{I,t} \cdot 1_{W_{I,t} > n^a/2} \right) > \frac{C n^\gamma}{\mathcal{D} n^{a}} \right]$$ To bound this final probability, we can directly bound the moment generating function of $W_{I,t} \cdot 1_{W_{I,t} > n^a/2}$: $$\mathbb{E}\left[\exp\left(W_{I,t} \cdot 1_{W_{I,t} > n^a/2} \right) \right] \leq 1 + \displaystyle \sum_{k > n^a/2} \frac{e^k}{k!} \leq 1 + \exp( - n^a) \, .$$ The first inequality follows because $W_{I,t}$ is a Poisson random variable of mean $1$ (or possibly less than $1$, if we pick the small $W_{I,t}$ in each $I$), while the second can deduced by using Stirling’s approximation and explicitly summing.
Applying a Chernoff strategy, we find that $$\mathbb{P}\left[\displaystyle \sum_{(I,t)} \left(W_{I,t} \cdot 1_{W_{I,t} > n^a/2} \right) > \frac{C n^\gamma}{\mathcal{D} n^{a}} \right] \leq \left(1 + \exp(-n^a) \right)^{m^d (\mathcal{D} +1)} \cdot \, \exp\left( - \frac{C n^\gamma}{\mathcal{D} n^{a}}\right) \, .$$ Using the standard approximation $(1+x) \leq e^{x}$, we find that the prefactor is bounded by $2$ for all $n$ sufficiently large. Combining all the earlier estimates, we find that $$\mathbb{P}[D] \leq \exp\left(- \frac{C n^{2\gamma}}{m^d \mathcal{D}^3 n^{4a}}\right) + 2 \exp\left( - \frac{C n^\gamma}{\mathcal{D} n^{a}}\right) \, .$$ The first term vanishes like $\exp(-n^{1 - 8a} f(n))$, whereas the second vanishes as $\exp( - n^{1-3a} f'(n))$, with $f$ and $f'$ varying more slowly than any polynomial or rational function. Thus, the first term dominates, and gives us the desired upper bound on the probability of $D$.
We now combine the technical lemmas of this section into a single, essential proposition:
\[HighProb\] Let $A$, $B$, $D$, and $L$ be defined as above. Then, in the conditional model, $\mathbb{P}(\cdot \mid L)$, the event $L \cap A^c \cap B^c \cap D^c$ occurs with high probability.
It is sufficient to show that the sum $\mathbb{P}(L^c | L) + \mathbb{P}(A | L) + \mathbb{P}(B | L) + \mathbb{P}(D | L)$ vanishes as $n$ grows. The first term is trivially equal to zero. To calculate the remaining three terms, we use the following obvious but very useful inequality: $$\mathbb{P}(A |L) = \frac{\mathbb{P}(A \cap L)}{\mathbb{P}(L)} \leq \frac{\mathbb{P}(A)}{\mathbb{P}(L)} \, .$$ By \[LLowerBound\] and \[BulkCardinality\], we know that $$\mathbb{P}(A |L) \leq \exp\left(q [\log (q/w) -1] + Cn^z \log n - C' n^{p/2 + a/2} \right) \, .$$ Since $a$ is positive, the negative term dominates, and the probability vanishes as $n$ grows. Referring to \[YIBound\], we use the same technique to find that $$\mathbb{P}(B |L) \leq \exp\left(Cn^z \log n - n^\beta/2 \right) \, .$$ Since $\beta > z$ by construction, this probability vanishes as well. \[ConcIneqSmall\] guarantees that, when $\mathcal{D}$ is vanishing, $$\mathbb{P}(D |L) \leq \exp\left(q [\log (q/w) -1] + Cn^z \log n - n^{2 \gamma - p - 8a}\right) \, .$$ The negative term grows as $n^{p- 12 a}$, while the positive one grows as $n^{p/2}$ (both up to polynomial factors). Since $p > a/25$, the exponent of the negative term is always greater, and thus the probability vanishes. Using the definition of $\mathcal{D}$, $m^d$, and \[ConcIneqLarge\], we can show that, for any $\mathcal{D}$ bounded below, $$\mathbb{P}(D |L) \leq \exp\left(q [\log (q/w) -1] + Cn^z \log n - n^{1 - 8a} \cdot f(n)\right) \, ,$$ where $f(n)$ grows or vanishes slower than any polynomial. As long as $p < 2 - 16 a$ (which is guaranteed by our assumptions on $p$), this probability vanishes, and the proof of the proposition is complete.
Showing Partial Localization
----------------------------
Surprisingly, the proposition that ends the previous subsection actually implies Theorem 3, in the sense that [*any*]{} configuration in $L \cap A^c \cap
B^c \cap D^c$ contains a maximal clique set such that the $X_I$ associated with every index is arbitrarily close to $(q/\tilde{\tau}_s)$. To prove this rather counterintuitive fact, we now embark on a technical proof which relies on careful asymptotics and convex analysis – and no probabilistic analysis.
For the remains of the proof of \[thm2\], all statements hold conditional on $L \cap A^c \cap B^c \cap D^c$, and, therefore, with high probability in the conditional model.
We now define some functionals of subsets of $T$, which will ease the notation. For any $W \subset T$, let $$Q(W) := \frac{2}{q^2}\sum_{I \in W} \biggl({X_I \choose 2} + \frac{1}{2} \displaystyle \sum_{J \in N_I\cap W} X_I X_J \biggr)\, .$$ This is the number of edges with both endpoints in $\mathfrak{U}(W)$, normalized by $q^2/2$. With this normalization, showing that most extra edges have both endpoints in $\mathfrak{U}(W)$ is equivalent to showing that $Q(W)$ is bounded below by something close to 1. Similarly, for two disjoint index sets $W$ and $W'$, we let $$Q(W,W') := \frac{2}{q^2} \sum_{I \in W} \displaystyle \sum_{J \in N_I \cap W'} X_I X_J\, .$$ This counts the edges between $W$ and $W'$ with the same normalization as above. Note that, for any $W \subset T$, $$|E_s| = \frac{q^2}{2} \big(Q(W) + Q(W,W^c) + Q(W^c)\big) \, .$$ Next, define $$V(W):= \frac{1}{q} \displaystyle \sum_{I \in W} X_I \, ,$$ the number of vertices in $W$. With this normalization, we can see that $V(W)^2 \geq Q(W)$. We also normalize the cardinality of sets by $\tilde{\tau}_s$, that is, for any $W \subset T$, $$h(W) := \frac{|W|}{\tilde{\tau}_s}\, .$$ As a final preliminary, we phrase Jensen’s inequality, an essential tool of this analysis, in terms of the quantities we defined:
\[lmm5\] For any $W \subset T$, $$\frac{1}{q} \displaystyle \sum_{I \in W} Y_I \geq V(W) \left[\log \left(\frac{q}{w}\right) + \log V(W) - \log h(W) - 1\right]$$
This is a direct application of Jensen’s inequality to the convex function $ Y_I = X_I \big(\log (X_I/\mathcal{D}) -1\big) + \mathcal{D}$.
We now consider the set of indices whose associated $X_I$’s are much larger than expected. The next lemma proves three bounds on this set:
\[mathfrakIBounds\] Define $\mathfrak{I}$ by $$\{ I \in T : X_I > \max \{ n^a, \mathcal{D} \cdot n^a \} \, .$$ Then, conditional on $L \cap A^c \cap B^c \cap D^c$, for any $\xi >0$ and $n$ sufficiently large (depending only on $\xi$), $$\begin{aligned}
\frac{1}{q} \sum_{I \in \mathfrak{I}} Y_I &\leq \log (q/w) - 1 + \xi\, ,\label{bound3}\\
V(\mathfrak{I}) &\leq C\, ,\label{bound2}\\
Q(\mathfrak{I}) &\geq 1 - \frac{\xi}{\log n}\, .\label{bound1}\end{aligned}$$
Note that the lower bound on $Q(\mathfrak{I})$ ensures that there are $\tilde{\delta} \tilde{\mu}_s$ edges - i.e. all the “excess" forced by the conditioning on $L$ - with both endpoints in $\mathfrak{U}(\mathfrak{I})$, up to lower order corrections. While the upper bound on $V(\mathfrak{I})$ is very loose, the upper bound on $\sum Y_I$ is both tight and essential to our analysis.
Defining $\mathfrak{I}$ as in the lemma above, we know that, thanks to $A^c$, $$\label{ICardBound}
|\mathfrak{I}| \leq n^{\alpha} \, .$$ With this cardinality bound, we know that $\mathfrak{I}$ falls within the definition of $B^c$, and therefore $$\sum_{I \in \mathfrak{I}} Y_I \leq q \left(\log (q/w) - 1\right) + n^{\beta}/2 \, .$$ Fix $\xi >0$. Dividing through by $q$ and noting that $n^{\beta}/q$ vanishes as $n$ grows, we can increase $n$ to ensure that holds, satisfying the first stipulation.
We now apply \[lmm5\] to this inequality. We deduce that $$V(\mathfrak{I}) \left[\log \left(\frac{q}{wh(\mathfrak{I})}\right) - 1\right] + V(\mathfrak{I}) \log V(\mathfrak{I}) \leq \log (q/w) - 1 + \xi\, .$$ If $V(\mathfrak{I}) \leq 1$, we have a (much better than needed!) bound on $V(\mathfrak{I})$. Otherwise, $V(\mathfrak{I}) \log V(\mathfrak{I})$ is positive, and we conclude that $$V(\mathfrak{I}) \leq \frac{\log (q/w) - 1 + C n^\beta/q }{\log (q/(wh(\mathfrak{I}))) - 1}\, .$$ By , and the definitions of the variables $q$, $w$ and $\alpha$, $$\frac{q}{w h(\mathfrak{I})} \geq C n^{a/2} \, ,$$ for some constant $C$; therefore the denominator grows at least as a constant multiple of $\log n$. Meanwhile, $q/w \leq C n^{1-p/2}$ for some constant $C$, meaning that $$V(\mathfrak{I}) \leq \frac{2(1-p/2) \log n }{(a/4) \log n } \leq \frac{8(1-p/2)}{a} \, .$$ We set the righthand side equal to $C$; since it is independent of $n$, the second stipulation has been proved.
For the final stipulation, we use the event $L \cap D^c$. By $D^c$, $$\label{ConcBound}
Q(\mathfrak{I}^c) - \mu \leq n^\gamma\, .$$ By the occurrence of $L$, $$\frac{q^2}{2} \big(Q(\mathfrak{I}) + Q(\mathfrak{I},\mathfrak{I}^c) + Q(\mathfrak{I}^c)\big) \geq (1 + \tilde{\delta}) \tilde{\mu}_s$$ Assuming $Q(\mathfrak{I}^c)$ takes on the maximal value allowed by , we deduce that $$\label{Qlb}
Q(\mathfrak{I}) + Q(\mathfrak{I},\mathfrak{I}^c) \geq 1 - \frac{2n^\gamma}{q^2} \, .$$ Next, we bound $Q(\mathfrak{I},\mathfrak{I}^c)$ by replacing each term in the sum over the neighborhood $N_I$ by the maximum allowable value: $$Q(\mathfrak{I},\mathfrak{I}^c) \leq \frac{2}{q^2} \sum_{I \in \mathfrak{I}} X_I \Big( \sum_{J \in N_I} \max_{J \in \mathfrak{I}^c} X_J\Big) \leq \frac{2
|N_I| M }{q} \Big(\frac{1}{q}\displaystyle \sum_{I \in \mathfrak{I}} X_I \Big)\, .$$ The parenthetical expression is $V(\mathfrak{I})$, which is at most a constant independent of $n$. By \[lmm3\], $|N_I|$ is also bounded above by a constant, and thus $$Q(\mathfrak{I},\mathfrak{I}^c) \leq \frac{C M}{q} \leq \frac{\xi}{2 \log n} \, ,$$ for sufficiently large values of $n$, since $M/q$ vanishes polynomially fast in $n$ for any value of $p$ allowed. Similarly, we can increase $n$ sufficiently to ensure that $(2 n^\gamma)/q$ is also bounded above by $\xi/(2 \log n)$ (which is possible thanks to the definition of $\gamma$). Substituting the bounds into completes the proof.
Improving the Linear Bound
--------------------------
Our goal is to produce an upper bound on the linear sum that matches the convex lower bound already in place for $\mathfrak{I}$. To do so, we find a subset of $\mathfrak{I}$ that has a similar number of “internal” edges, but which only involves approximately $q$ vertices. Using the convex upper bound from above, we construct such a subset in the simplest way possible: taking the largest elements of $\mathfrak{I}$, and stopping when the total number of vertices gets close to q. However, although we hope that the cardinality of this set has a uniform upper bound in $n$, our argument is too rough to prove such a statement. Therefore, we ensure that the bounds get tighter as the size of the set increases; later these bounds will allow us to enforce an upper bound on the number of indices we are concerned with.
\[lmm6\] Let $\mathfrak{I}$ be the set defined in equation \[mathfrakIBounds\]. Then there exists $\mathfrak{T} \subset \mathfrak{I}$ such that $$\begin{aligned}
Q(\mathfrak{T}) &\geq 1 - \psi(\mathfrak{T})\,\end{aligned}$$ and $$\begin{aligned}
1 - \frac{2\xi}{\log n} &\leq V(\mathfrak{T}) \leq 1 + \phi(\mathfrak{T})\end{aligned}$$ where $$\phi(\mathfrak{T}) = \min \left\{\xi/3, \, \frac{2}{|\mathfrak{T}|} - \frac{2 \xi}{\log n}\right\}$$ and $$\psi(\mathfrak{T}) = \min \left\{ \xi,\, \frac{C'}{|\mathfrak{T}|} \right\}$$ for some $C'$ independent of $n$.
Since $Q(\mathfrak{I}) \geq 1 - \xi/\log n$ by , we know that $$V(\mathfrak{I}) \geq \sqrt{1 - \xi/\log n} \geq 1 - \xi/ \log n$$ for all sufficiently large $n$. Now, we order the elements of $\mathfrak{I}$ by size; for the purposes of this proof, we label the $X_I$’s by integers: let $X_1$ the largest $X_I$ among all $I\in \mathfrak{I}$, $X_2$ be the second largest, and so on. Pick an integer $k$, and define $\mathfrak{I}_k$ to be the first $k$ terms of $\mathfrak{I}$ with respect to this order. Finally, define $$\label{mtdef}
\mathfrak{T} := \text{ The first $\mathfrak{I}_k$ such that $V(\mathfrak{I}_k) > 1 - 2 \xi/\log n$}\, .$$ that is, the smallest subset of $\mathfrak{I}$ that has at least $q(1 - 2 \xi/ \log n)$ vertices of $\chi$ in its associated $A_I$’s. By minimality, $$V\left(\mathfrak{T}\backslash \{X_{|\mathfrak{T}|}\} \right) \leq 1 - \frac{2 \xi}{\log n}$$ Furthermore, $X_{|\mathfrak{T}|}$ must be the minimal element of $\mathfrak{T}$, and therefore, $$\frac{X_{|\mathfrak{T}|}}{q} \leq \frac{V(\mathfrak{T})}{|\mathfrak{T}|}\, .$$ Substituting these two inequalities into the identity $V(\mathfrak{T}) = V(\mathfrak{T}\backslash \{X_{|\mathfrak{T}|}\}) + X_{|\mathfrak{T}|}/q$ and solving for $V(\mathfrak{T})$ gives $$V( \mathfrak{T} ) \leq 1 + \frac{2}{|\mathfrak{T}|} - \frac{2 \xi}{\log n}\, .$$ Note that this inequality may be derived from the previous one only if $|\mathfrak{T}|$ is larger than a universal constant. That, however, is not a problem; Corollary \[newcor\] at the end of the proof shows that $|\mathfrak{T}|$ must be large enough, simply using the definition of $\mathfrak{T}$ and Lemma \[lmm5\].
This establishes one half of the desired upper bound on $V(\mathfrak{T})$. It remains to show that $V(\mathfrak{T})$ is bounded by $\xi$ when $|\mathfrak{T}| < 6/\xi$. Assume that that bound, and apply Lemma \[lmm5\] to $\mathfrak{T}$: $$\begin{aligned}
&V(\mathfrak{T}) \left[\log \left(\frac{q}{w}\right) - \log h(\mathfrak{T})- 1\right] + V(\mathfrak{T}) \log V(\mathfrak{T}) \\
&\leq \frac{1}{q} \displaystyle \sum_{I \in \mathfrak{T}} Y_I \, \leq \, \, \log(q/w) - 1 + \xi \, ,\end{aligned}$$ where the last upper bound follows by and the fact that $\mathfrak{T} \subset \mathfrak{I}$. From the above inequality it follows that either $V(\mathfrak{T}) \leq 1$, or $$V(\mathfrak{T}) \leq \frac{\log(q/w) - 1 + \xi}{\log(q/w) - \log h(\mathfrak{T}) - 1}\, .$$ The assumed upper bound on $|\mathfrak{T}|$ implies that $h(\mathfrak{T})$ is uniformly bounded; since $|\log (q/w)|\to \infty$ as $n\to \infty$, we conclude the second upper bound on $V(\mathfrak{T})$.
Next, we need to show that $Q(\mathfrak{T})$ is large. In the remaining part of this proof, $\mathfrak{T}^c$ will denote the complement of $\mathfrak{T}$ in $\mathfrak{I}$, and not in all of $T$. It is sufficient to bound $$Q(\mathfrak{T}, \mathfrak{T}^c) + Q(\mathfrak{T}^c)$$ from above, by partitioning the sum that defines $Q(\mathfrak{I})$, as in the previous section. The maximum element in $\mathfrak{T}^c$ cannot exceed $(q
V(\mathfrak{T}))/|\mathfrak{T}|$, as it must be smaller than the minimum of the elements in $\mathfrak{T}$, by definition. Therefore, for any $I \in
\mathfrak{I}$, $$\displaystyle \sum_{J \in N_I\cap \mathfrak{T}^c} X_J \leq |N_I|\max_{J \in \mathfrak{T}} X_J\leq |N_I| \cdot \left( \frac{q V(\mathfrak{T})}{|\mathfrak{T}|}\right) \, .$$ Expanding the definition of both terms, we see that $$\label{qq}
Q(\mathfrak{T}, \mathfrak{T}^c) + Q(\mathfrak{T}^c) \leq \frac{4}{q^2} \displaystyle \sum_{I \in \mathfrak{I}} X_I \displaystyle \sum_{J \in N_I\cap
\mathfrak{T}^c} X_J \leq [V(\mathfrak{I})V(\mathfrak{T})] \frac{|N_I|}{|\mathfrak{T}|}\, .$$ Setting $C' = 2 C |N_I|$, where $C$ is the constant from \[mathfrakIBounds\] and assuming $|\mathfrak{T}| > C'/\xi$, we see that this $$Q(\mathfrak{T}, \mathfrak{T}^c) + Q(\mathfrak{T}^c) \leq \psi(\mathfrak{T}) \, ,$$ noting that $V(\mathfrak{T}) \leq 1 + \phi(\mathfrak{T}) \leq 2$.
Finally, assume that $|\mathfrak{T}| \leq C'/\xi$. We go back to the upper bound on $Y_I$’s in $\mathfrak{I}$: $$\frac{1}{q}\biggl(\displaystyle \sum_{I \in \mathfrak{T}} Y_I + \sum_{J \in \mathfrak{T}^c} Y_J\biggr) \leq \log(q/w) - 1 + \xi\, .$$ We apply Lemma \[lmm5\] to the term involving $\mathfrak{T}$ in the above inequality to conclude that $$\begin{aligned}
&V(\mathfrak{T}) \left[\log\left(\frac{q}{w}\right) + \log V(\mathfrak{T}) - \log h(\mathfrak{T}) -1 \right] + \frac{1}{q} \sum_{J \in \mathfrak{T}^c}
Y_J \\
&\leq \log (q/w) - 1 + \xi\, .\end{aligned}$$ Next, we substitute the lower bound on $V(\mathfrak{T})$ (from definition). After some algebraic manipulation, we conclude that $$\begin{aligned}
\frac{1}{q} \sum_{J \in \mathfrak{T}^c} Y_J &\leq \xi + \frac{3 \xi}{\log n} + \log h(\mathfrak{T}) \\
&\qquad + \frac{2 \xi}{\log n} \left[\log\left(\frac{q}{w}\right) - \frac{3 \xi}{\log n} - \log h(\mathfrak{T}) -1\right]\, .\end{aligned}$$ (To get this, we bounded $\log(1 - 2\xi/\log n)$ from below by $-3\xi/\log n$, an estimate that holds for all sufficiently large $n$.) Thanks to the assumed upper bound on $|\mathfrak{T}|$, the right-hand size can be bounded above by $2 \log(1/\xi)$ for all sufficiently small $\xi$ and sufficiently large $n$, as all terms except $h(\mathfrak{T})$ are bounded above by a constant multiple of $\xi$.
With this upper bound, we can now apply Lemma \[lmm5\] to $\mathfrak{T}^c$ to conclude that $$V(\mathfrak{T}^c) \left[\log\left(\frac{q}{w h(\mathfrak{T}^c)} \right)-1\right] + V(\mathfrak{T}^c) \log V(\mathfrak{T}^c) \leq 2 \log (1/ \xi)\, .$$ Since the absolute minimum of the function $x \log x$ is $-1/e$, we can replace $V(\mathfrak{T}^c) \log V(\mathfrak{T}^c)$ by this minimum and maintain the lower bound. We also know that $w h(\mathfrak{T}^c)$ is bounded above by $\mathcal{D} n^{\alpha}$ (since $\mathfrak{T}^c \subset \mathfrak{I}$ and the event $A^c$ is assumed to have happened). Recalling the definition of $\alpha$, we now arrive at the inequality $$V(\mathfrak{T}^c) \leq \frac{2 \log(1/\xi) + 1/e}{\log(q/(\mathcal{D} n^{\alpha})) -1} \leq \frac{3 \log(1/\xi)}{a/2 \log n} \leq \frac{\xi}{C}\, ,$$ where $C$ is, again, the constant from \[mathfrakIBounds\]. The final inequality follows since the numerator is bounded in $n$, while the denominator grows, and therefore the fraction can be made smaller than any constant. From here, a bound on $Q(\mathfrak{T},\mathfrak{T}^c) + Q(\mathfrak{T}^c)$ is easy: For any $I \in \mathfrak{I}$, $$\frac{1}{q} \displaystyle \sum_{J \in N_I\cap \mathfrak{T}^c} X_J \leq V(\mathfrak{T}^c)\, .$$ Therefore, $$Q(\mathfrak{T}, \mathfrak{T}^c) + Q(\mathfrak{T}^c) \leq \frac{4}{q^2} \displaystyle \sum_{I \in \mathfrak{I}} X_I \displaystyle \sum_{J \in N_I\cap
\mathfrak{T}^c} X_J \leq \left(\frac{\xi}{C}\right) V(\mathfrak{I}) \leq \xi$$ Completing the proof of the lemma.
\[newcor\] There exists $C$ independent of n such that $$|\mathfrak{T}| > \tilde{\tau}_s\, .$$
Since $\mathfrak{T} \subset \mathfrak{I}$, applies to $\mathfrak{T}$ as well. We now apply Lemma \[lmm5\] to this set, and conclude that $$V(\mathfrak{T}) \left[\log \left(\frac{q}{w}\right) + \log V(\mathfrak{T}) - \log h(\mathfrak{T}) - 1\right] \leq \log \left(\frac{q}{w}\right) + \xi -1$$ From the definition of $\mathfrak{T}$ in the proof of Lemma \[lmm6\], we have a lower bound on $V(\mathfrak{T})$. Noting that $$\log\left(1 - \frac{2 \xi}{\log n}\right) \geq \frac{-4 \xi}{\log n}$$ for all sufficiently large $n$, we can conclude that
$$- \left(1 - \frac{2 \xi}{\log n} \right) \log(h (\mathfrak{T})) \leq \xi + \frac{2 \xi \log (q/w)}{\log n} + \frac{C \xi}{\log n}$$ We recall that $q/w \leq C n^{1-p/2}$, and therefore there exists a constant $C$ such that $$- \log(h (\mathfrak{T})) \leq C \xi\, .$$ Inverting the negative logarithm gives $$h(\mathfrak{T}) \geq \exp(-C \xi) \geq 1 - C \xi$$ If we decrease $\xi$ enough to ensure that $C \xi < \tilde{\tau}_s/2$ (which is possible, as $\tilde{\tau}_s$ is independent of $n$), we find that $$|\mathfrak{T}| \geq \tau_s - 1/2 \, .$$ Since the cardinality is an integer, we get the desired bound.
The estimates of \[lmm6\] imply a that $V(\mathfrak{T})$ satisfies nearly matching upper and lower bounds; we now exploit this property to bound an algebraic quantity that has a nice geometric interpretation:
\[PIbound\] For any $W \subset T$, define $$P_I(W) := \frac{1}{q} \sum_{J \in W,\, d(I,J) >s} X_J \, .$$ Note that the sum is over indices whose distance [*exceeds*]{} $s$. Then $$\frac{1}{q} \sum_{I \in \mathfrak{T}} X_I P_I(\mathfrak{T}) \leq 3 \phi(\mathfrak{T})+ \psi(\mathfrak{T}) \, ,$$ where $\mathfrak{T}$, $\phi(\cdot)$ and $\psi(\cdot)$ are defined as in \[lmm6\].
First, we consider the quantity $[V(\mathfrak{T})]^2 - Q(\mathfrak{T})$. By expanding the product and cancelling, we can deduce that $$[V(\mathfrak{T})]^2 - Q(\mathfrak{T}) = \frac{2}{q} \sum_{I \in \mathfrak{T}} X_I P_I(\mathfrak{T}) + \frac{V(\mathfrak{T})}{q} \, .$$ Note that $$X_I^2 - 2{X_I \choose 2} = X_I \, ,$$ which accounts for the additive factor of $V(\mathfrak{T})/q$. By \[lmm6\], we conclude that $$\frac{1}{q} \sum_{I \in \mathfrak{T}} X_I P_I(\mathfrak{T}) \leq \left([V(\mathfrak{T})]^2 - Q(\mathfrak{T})\right) \leq (1 + \phi(\mathfrak{T}))^2 - (1
- \psi(\mathfrak{T})) \leq 3 \phi(\mathfrak{T}) + \psi(\mathfrak{T}) \, .$$
The quantity $\sum X_I P_I(\mathfrak{T})$ is commensurate with the number of edges in the complement graph of $G_s(\chi,r)$, restricted to vertices in $\mathfrak{U}(\mathfrak{T})$. We wish to show that, whenever it is small, the graph restricted to $\mathfrak{T}$ is not far from a clique. This will be the goal of the next section.
Removing Lower Order Terms
--------------------------
Before proceeding, consider the case $\xi =0$. In this case, both $\psi$ and $\phi$ vanish. Since all terms in lefthand sum are strictly positive (in fact, $\mathfrak{T} \subset \mathfrak{I}$, so every $X_I$ we consider is at least $M$), $P_I(\mathfrak{T})$ would be forced to vanish for each $I \in
\mathfrak{T}$. This would imply that the diameter of $\mathfrak{T}$ is at most $s$. By \[newcor\], the cardinality of the set is at least $\tilde{\tau}_s$ – and thus $\mathfrak{T}$ would be a maximal clique set!
Since $\xi$ is strictly positive, we cannot apply this argument to the set $\mathfrak{T}$. The next lemma produces a subset of $\mathfrak{T}$ with slightly stronger bounds; at the end of the section, we will apply an argument very similar to the one described above to prove that that set is a maximal clique set.
\[lmm7\] Let $\mathfrak{T}$ be as above, and define $\mathfrak{P}$ such that $$\mathfrak{P} := \left\{I \in \mathfrak{T} : X_I > \frac{\xi^{1/4} q}{\tilde{\tau}_s} \right\}$$ Then, for sufficiently small $\xi >0$ and sufficiently large $n$, $$1 - \xi^{1/5} \leq V(\mathfrak{P}) \leq 1 + \phi(\mathfrak{T})$$ and $$\frac{1}{q} \sum_{I \in \mathfrak{P}} Y_I \leq V(\mathfrak{P}) (\log(q/w) - 1) + 2\xi^{1/10}\, .$$
We split $\mathfrak{T}$ into three sets: $\mathfrak{P}$ is defined as above, while the rest of the indices are split as follows: $$\mathfrak{L}_1 := \biggl\{ I \in \mathfrak{T}: X_I \leq \frac{\xi^{1/4} q}{\tilde{\tau}_s \log |\mathfrak{T}|}\biggr\}$$ and $$\mathfrak{L}_2 := \biggl\{ I \in \mathfrak{T}: \frac{\xi^{1/4} q}{\tilde{\tau}_s \log |\mathfrak{T}|} \leq X_I \leq \frac{\xi^{1/4} q}{\tilde{\tau}_s }
\biggr\}\, .$$ Since the three sets partition $\mathfrak{T}$, we can prove the lemma as with sufficiently good upper bounds on $V(\mathfrak{L}_i)$ and lower bounds on the sum of the $Y_I$’s in both sets.
To bound $V(\mathfrak{L}_1)$, we first need to bound $P_I(\mathfrak{L}_1)$ from below. The worst case scenario is that the distance restriction removes the $|N_I|$ largest elements of $\mathfrak{L}_1$. Therefore, $$P_I(\mathfrak{L}_1) \geq V(\mathfrak{L}_1) - \frac{1}{q}|N_I|\left( \max_{J \in \mathfrak{L}_1} X_J \right) \geq V(\mathfrak{L}_1) - \frac{|N_I|
\xi^{1/4}}{\tilde{\tau}_s \log |\mathfrak{T}|}\, .$$ Since $P_I(W) \leq P_I(W')$ whenever $W \subset W'$, we see that \[PIbound\] implies that $$\frac{1}{q} \sum_{I \in \mathfrak{T}} X_I P_I(\mathfrak{L}_1) \leq 3 \phi(\mathfrak{T}) + \psi(\mathfrak{T})\, .$$ Replacing $P_I(\mathfrak{L}_1)$ with its minimum and recalling that $|N_I|/\tilde{\tau}_s$ is uniformly upper bounded (by Lemma \[lmm3\]), we see that $$\biggl(V(\mathfrak{L}_1) - \frac{C \xi^{1/4}}{\log |\mathfrak{T}|} \biggr) V(\mathfrak{T}) \leq 3 \phi(\mathfrak{T}) + \psi(\mathfrak{T})\, .$$ Using the (very suboptimal) lower bound of $1/2$ for $V(\mathfrak{T})$, we conclude that $$\label{vl1}
V(\mathfrak{L}_1) \leq 6\phi(\mathfrak{T}) + 2 \psi(\mathfrak{T}) + \frac{C \xi^{1/4}}{\log |\mathfrak{T}|}\, .$$ for some $C$ independent of $n$. Repeating this analysis with $\mathfrak{L}_2$ yields the inequality $$V(\mathfrak{L}_2) \leq 6\phi(\mathfrak{T}) + 2 \psi(\mathfrak{T}) + C \xi^{1/4}\, .$$ Since both $\phi$ and $\psi$ are bounded above by $\xi$, we get $$\label{vbounds}
\max\{V(\mathfrak{L}_1), \, V(\mathfrak{L}_2)\} \leq \xi^{1/5}/3.$$ Now, $$V(\mathfrak{P}) = V(\mathfrak{T}) - V(\mathfrak{L}_1) - V(\mathfrak{L}_2) \geq 1 - \frac{2\xi}{\log n} - \frac{2 \xi^{1/5}}{3} \geq 1 - \xi^{1/5}$$ This establishes the lower bound on $V(\mathfrak{P})$. The upper bound follows trivially from $\mathfrak{P} \subset \mathfrak{T}$.
Finally, we need to improve the upper bound on the $Y_I$’s associated with $\mathfrak{P}$. Not surprisingly, we will use Jensen’s inequality to get lower bounds on the $Y_I$ associated with $\mathfrak{L}_1$ and $\mathfrak{L}_2$. By inclusion, we know that $$\frac{1}{q} \sum_{I \in \mathfrak{P}} Y_I \leq \left(\log(q/w) -1 + \xi\right) - \displaystyle \sum_{I \in \mathfrak{L}_1 \cup \mathfrak{L_2} } Y_I$$ Suppose that $$\label{ineq32}
\frac{1}{q}\sum_{I \in \mathfrak{L}_i} Y_I \geq V(\mathfrak{L}_i)\left(\log(q/w) -1\right) - \xi^{1/10}/2$$ for $i=1,2$. The inequality $$1 - V(\mathfrak{L}_1) - V(\mathfrak{L}_2) \leq V(\mathfrak{P}) + (2\xi)/\log n$$ follows from partitioning $\mathfrak{T}$ into its three constituent sets, and the lower bound on $V(\mathfrak{T})$ from Lemma \[lmm6\]. Substituting the two inequalities into the earlier statement gives $$\begin{aligned}
\frac{1}{q}\sum_{I \in \mathfrak{P}} Y_I &\leq V(\mathfrak{P})\left(\log(q/w) -1 \right) + \xi + \frac{2 \xi}{\log n} \left(\log(q/w) -1\right) +
\xi^{1/10}\\
&\leq V(\mathfrak{P})\left(\log(q/w) -1 \right) + 2 \xi^{1/10}\, .\end{aligned}$$ using the fact that $\log(q/w)$ is bounded above by a constant multiple $\log n$. Since this is the required inequality in the statement of the lemma, we will be done if we can prove the inequality .
By applying Jensen’s inequality \[lmm5\] to $\mathfrak{L}_i$, we know that $$\label{JensenLi}
\frac{1}{q}\sum_{I \in \mathfrak{L}_i } Y_I \geq V(\mathfrak{L}_i) \left(\log(q/w) -1 \right) + V(\mathfrak{L}_i) \log V(\mathfrak{L}_i) -
V(\mathfrak{L}_i) \log h(\mathfrak{L}_i) \, .$$ We have already shown in that $V(\mathfrak{L}_i)$ is small. Now, the function $\xi^{a} \log \xi$ is bounded below $- \xi^{b}/4$ for any $b
< a$ and $\xi$ sufficiently close to zero, and therefore, $$V(\mathfrak{L}_i) \log V(\mathfrak{L}_i) \geq - \xi^{1/10}/4 \, .$$ For the final term, we must employ a different strategy for $\mathfrak{L}_1$ and $\mathfrak{L}_2$. If $|\mathfrak{L}_1| \leq C/\xi$, then $-V(\mathfrak{L}_1) \log h(\mathfrak{L}_1) > C \xi^{1/5} \log \xi$. Repeating the argument from before, the final term can be bounded below by $-\xi^{1/10}/4$, as required. Now suppose $|\mathfrak{L}_1| > C/\xi$. In this range, both $\phi(\mathfrak{T})$ and $\psi(\mathfrak{T})$ are bounded above by $C/|\mathfrak{T}|$. Thus, implies that $$V(\mathfrak{L}_1) \leq \frac{\xi^{1/5}}{\log |\mathfrak{T}|} \, .$$ Since $\mathfrak{L}_1 \subset \mathfrak{T}$, we deduce that $h(\mathfrak{L_1}) \leq C |\mathfrak{T}|$, and $$- V(\mathfrak{L}_i) \log h(\mathfrak{L}_i) > - C \xi^{1/5} > -\xi^{1/10}/4 \, .$$ This completes the lemma for $\mathfrak{L}_1$.
For $\mathfrak{L}_2$, we claim that $|\mathfrak{L}_2| < C/\xi^{1/4}$. To see this, first assume that $P_I(\mathfrak{L}_2) = 0$ for some $I$. This implies that $|\mathfrak{L}_2| \leq |N_I|$, and $h(\mathfrak{L}_2)$ is bounded above by a constant by \[lmm3\]. Now, assume that $P_I(\mathfrak{L}_2)$ is nonzero for all $I$. Then we can minimize the value of $P_I$ by assuming that the sum only includes the minimal elements of $\mathfrak{L}_2$: $$P_{I}(\mathfrak{L}_2) \geq \left(\frac{|\mathfrak{L}_2| - |N_I|}{q} \right) \cdot \left(\min_{J \in \mathfrak{L}_2} X_J\right) \geq \frac{C \xi^{1/4}
|\mathfrak{L}_2| - |N_I|}{\log |\mathfrak{T}|} \, ,$$ for some $C$ independent of $n$, where the second inequality follows from the definition of $\mathfrak{L}_2$ (and using \[lmm3\] to push the factors of $\tilde{\tau}_s$ into $C$). Multiplying by $X_I/q$, summing over $I$’s in $\mathfrak{T}$, and using \[PIbound\], we see that $$\begin{aligned}
V(\mathfrak{T}) \cdot \left(\frac{C \xi^{1/4} |\mathfrak{L}_2| - |N_I|}{\log |\mathfrak{T}|} \right)& \leq \frac{1}{q}\sum_{I \in \mathfrak{T}} X_I
P_I(\mathfrak{L}_1) \\ & \leq 3 \phi(\mathfrak{T}) + \psi(\mathfrak{T}) \, .\end{aligned}$$ Solving for $|\mathfrak{L}_2|$, we deduce that $$|\mathfrak{L}_2| \leq |N_I| + \frac{C \log |\mathfrak{T}| (\phi(\mathfrak{T}) + \psi(\mathfrak{T}))}{\xi^ {1/4}} \, ,$$ where we use the upper bound on $V(\mathfrak{T})$. Since both $\phi(\mathfrak{T})$ and $\psi(\mathfrak{T}))$ are bounded above by $2/|\mathfrak{T}|$, the numerator is at most a constant independent of $n$. Therefore, $|\mathfrak{L}_2| < C/\xi^{1/4}$. To complete the proof, we simply repeat the argument used in the case $|\mathfrak{L}_1| \leq C/\xi$.
As promised, we now show that $\mathfrak{P}$ has the desired geometric properties:
\[MaximalClique\] The set $\mathfrak{P}$ is a maximal clique set.
Assume that there exists a pair of indices $I,J \in \mathfrak{P}$ such that $d(I,J) >s$. By the definition of $\mathfrak{P}$, $$\frac{X_I X_J}{q^2} \geq \frac{\xi^{1/2}}{\tilde{\tau}_s^2}\, .$$ By \[PIbound\], we know this cannot exceed $3 \phi(\mathfrak{T}) + \psi(\mathfrak{T})$, which is bounded above by $4\xi$. This forces $$\frac{\xi^{1/2}}{\tilde{\tau}_s^2} \leq 4 \xi\, .$$ However, the above inequality can easily be rendered false by choosing $\xi$ so small that $\xi \leq (2\tilde{\tau}_s)^{-4}$. This can be done, because by Lemma \[lmm3\], $\tilde{\tau}_s$ is uniformly bounded by a constant. With such a choice of $\xi$, the diameter of $\mathfrak{P}$ is bounded above by $s$.
Combining Lemma \[lmm5\] and \[lmm7\] gives $$\begin{aligned}
&V(\mathfrak{P}) \left(\log(q/w) - 1 \right) + V(\mathfrak{P}) \big(\log V(\mathfrak{P}) - \log h(\mathfrak{P})\big) \\
&\leq V(\mathfrak{P}) \left(\log(q/w) - 1 \right) + 2 \xi^{1/10}\, ,\end{aligned}$$ and therefore $$h(\mathfrak{P}) \geq V(\mathfrak{P}) \exp \biggl(- \frac{2 \xi^{1/10}}{V(\mathfrak{P})}\biggr) \geq V(\mathfrak{P}) \biggl(1 - \frac{2
\xi^{1/10}}{V(\mathfrak{P})}\biggr)$$ using the standard estimate $e^{-x} \ge 1-x$. Combining this with the lower bound on $V(\mathfrak{P})$ from Lemma \[lmm7\] forces $$|\mathfrak{P}| = \tilde{\tau}_s h(\mathfrak{P}) \geq \tilde{\tau}_s (1 - 3 \xi^{1/10})$$ Again, since $\xi$ is under our control, we set $\xi < (2 \tilde{\tau}_s)^{-10}$. This implies that $$|\mathfrak{P}| \geq \tilde{\tau}_s - \frac{3}{2^{10}}$$ Cardinality must be an integer, and therefore $|\mathfrak{P}| \geq \tilde{\tau}_s$. Therefore, we have shown that $\mathfrak{P}$ is a set of diameter at most $s$, with at least $\tilde{\tau}_s$ elements. But, by definition, $\tilde{\tau}_s$ is the largest possible cardinality of any set of diameter $s$. Thus, $\mathfrak{P}$ must have diameter $\textit{exactly}$ $s$ — otherwise, consider the union of $\mathfrak{P}$ with one of the indices in $T$ at distance $1$ from $\mathfrak{P}$. By the triangle inequality, the diameter of this set is at most $s$, but its cardinality is $\tilde{\tau}_s +1$, contradicting maximality. Thus, $\mathfrak{P}$ is a maximal clique set.
Convex Analysis
---------------
We are nearly done with the proof: all that remains is to show that the elements of $\mathfrak{P}$ are close to $q/\tilde{\tau}_s$, and to improve the upper bound for maximal clique sets far away from $\mathfrak{P}$.
\[QuantitatveJensens\] Let $\mathfrak{P}$ be as above. Then, for any $I$ in $\mathfrak{P}$, $$\left|\frac{X_I \tilde{\tau}_s}{q} - 1 \right| < \xi^{1/40} \, .$$
We consider the Taylor’s expansion of $Y_I$ around the value $q/\tilde{\tau}_s$. Explicitly, we let $f(x) = x (\log(x/\mathcal{D}) -1) + \mathcal{D})$, and by Taylor’s theorem, $$Y_I = f\biggl(\frac{q}{\tilde{\tau}_s}\biggr) + f'\biggl(\frac{q}{\tilde{\tau}_s}\biggr) \biggl(X_I - \frac{q}{\tilde{\tau}_s}\biggr) +
\frac{f''(L(X_I))}{2} \biggl(X_I - \frac{q}{\tilde{\tau}_s}\biggr)^2$$ where $L(X_I)$ is some number between $X_I$ and $q/\tilde{\tau}_s$. Differentiating $f(x)$ explicitly and simplifying algebraically,, we see that $$Y_I = \mathcal{D} - \frac{q}{\tilde{\tau}_s} + X_I \log(q/w) + \frac{1}{2L(X_I)} \biggl(X_I - \frac{q}{\tilde{\tau}_s}\biggr)^2\, .$$ Next, we sum over $\mathfrak{P}$ and use the upper bound from Lemma \[lmm7\] to deduce that $$\begin{aligned}
&\frac{1}{q}\displaystyle \sum_{I \in \mathfrak{P}}\biggl[\mathcal{D} - \frac{q}{\tilde{\tau}_s} + X_I\log(q/w) + \frac{1}{2L(X_I)} \biggl(X_I -
\frac{q}{\tilde{\tau}_s}\biggr)^2 \biggr] \\
&\leq V(\mathfrak{P}) (\log(q/w) -1) + 2 \xi^{1/10}\, .\end{aligned}$$ Recalling that $\mathfrak{P}$ has exactly $\tilde{\tau}_s$ elements, a bit of algebraic manipulation yields $$\label{lxi}
\frac{1}{q}\sum_{I \in \mathfrak{P}} \frac{1}{2L(X_I)} \biggl(X_I - \frac{q}{\tilde{\tau}_s}\biggr)^2 \leq 1 - V(\mathfrak{P}) + 2\xi^{1/10} \leq 3
\xi^{1/10} \, ,$$ where the final inequality follows thanks to the lower bound on $V(\mathfrak{P})$ from \[lmm7\].
Now, define $$\mathfrak{W}_1:= \left\{I \in \mathfrak{P}: X_I \geq (1 + \xi^{1/40}) q/\tilde{\tau}_s\right\}$$ and $$\mathfrak{W}_2:= \left\{I \in \mathfrak{P}: X_I \leq (1 - \xi^{1/40}) q/\tilde{\tau}_s \right\}\, .$$ On $\mathfrak{W}_1$, the function $1/L(X_I)$ is bounded below by $1/X_I$. Thus, implies that $$\frac{|\mathfrak{W}_1|}{q} \min_{I \in \mathfrak{W}_1} \left( \frac{1}{2X_I} \biggl(X_I - \frac{q}{\tilde{\tau}_s}\biggr)^2\right) \leq 3
\xi^{1/10} \, .$$ The function $x \mapsto (x - q/\tilde{\tau}_s)^2/(2x)$ is strictly increasing on the interval $[(q/\tilde{\tau}_s)(1 +\xi^{1/40}), \infty)$, and thus, $$|\mathfrak{W}_1| \cdot \left( \frac{\xi^{1/20}}{2\tilde{\tau}_s \cdot (1 + \xi^{1/40})} \right) \leq 4 \xi^{1/10} \, .$$ Reducing $\xi$ such that $\xi < (10 \tilde{\tau}_s)^{-20}$ implies that $|\mathfrak{W}_1|$ is bounded above by $4/5$; since the cardinality is an integer, the set is empty. For $\mathfrak{W}_2$, we can bound $1/L(X_I)$ from below by $\tilde{\tau}_s/q$. Thus, $$\frac{|\mathfrak{W}_2|}{q} \min_{I \in I \in \mathfrak{W}_2} \left( \frac{\tilde{\tau}_s}{2q} \biggl(X_I - \frac{q}{\tilde{\tau}_s}\biggr)^2\right) \geq
\frac{|\mathfrak{W}_2| \xi^{1/20} }{2 \tilde{\tau}_s} \, .$$ Substituting this into and reducing $\xi$ sufficiently completes the proof.
We have completed the proof of the difficult assertion in \[thm2\]; all that is left is to ensure the second stipulation holds.
Set $\xi = \min\{ \tilde{\epsilon}^{40},1\}$. By the sequence of assertions \[HighProb\], \[mathfrakIBounds\], \[lmm6\], \[lmm7\], \[MaximalClique\] and \[QuantitatveJensens\], we produce a maximal clique set $\mathfrak{P}$ such that $$\left|\frac{X_I \tilde{\tau}_s}{q} - 1 \right| < \tilde{\epsilon} \, ,$$ proving the first stipulation of \[thm2\].
For the second stipulation, we must show that all elements of $\mathfrak{P}^c$ are small. If $\mathfrak{T} \backslash \mathfrak{P}$ is nonempty, the largest element in that set must satisfy $X_I < \xi^{1/4} q/\tau_s$. Furthermore, every element in $\mathfrak{I}\backslash \mathfrak{T}$ is smaller than every element in $\mathfrak{T}$, implying this bound is inherited by all elements of $\mathfrak{I} \backslash \mathfrak{P}$. This is enough to pass the upper bound to all remaining $X_I$’s, as required.
We are left with the scenario in which $\mathfrak{P} = \mathfrak{T}$. Formally, it is still possible that the largest $X_I$ with $I \in \mathfrak{I}\backslash \mathfrak{T}$ is of order $q$. However, we now know that $h(\mathfrak{T}) = 1$. By Lemma \[lmm5\] and the lower bound in Lemma \[lmm6\], we deduce that $$\begin{aligned}
\frac{1}{q} \displaystyle \sum_{I \in \mathfrak{T}} Y_I &\geq V(\mathfrak{T}) \left(\log(q/w) +\log(V(\mathfrak{T})) - 1 \right) \\
&\geq \left(1 - \frac{2 \xi}{ \log n}\right) \left(\log(q/w) - 1 \right)\, .\end{aligned}$$ Therefore, we find that $$\begin{aligned}
\frac{1}{q} \displaystyle \sum_{I \in \mathfrak{I} \backslash \mathfrak{T}} Y_I &\leq \left[\log(q/w) -1 + \xi \right] - \left(1 - \frac{2 \xi}{\log n}\right) \left[\log(q/w) - 1\right] \\
& \leq \frac{ 4\xi \log(q/w)}{\log n} \\
& \leq C \xi\, .\end{aligned}$$ As before, the final inequality follows because $\log (q/w) \leq C \log n$ for some $C$.
If there is some $I^* \in \mathfrak{I} \backslash \mathfrak{T}$ with $X_{I^*} > \xi q$, we find that $$\begin{aligned}
Y_{I^*} & = X_{I^*} \left(\log(X_{I^*}/\mathcal{D}) - 1\right) +\mathcal{D} \\
&\geq \xi q \left(\log(q/\mathcal{D}) - 1 \right) \, ,\end{aligned}$$ where we ignore several positive terms in the final inequality. If we divide through by $q$, we find that this expression still grows with $n$, whereas the earlier upper bound is uniformly bounded. This is a contradiction, and therefore every $X_I \in \mathfrak{I} \backslash \mathfrak{T}$ is bounded by $\xi q$. This completes the proof.
Proof of the Upper Tail Large Deviation Principle
=================================================
As usual, we write $\mu$, $\tau$, and $r$ instead of $\mu_n$, $\tau_n$ and $r_n$. We now prove Theorem \[thmLDP\], which claims that the function $$I(x) := \left(\frac{2 -p}{2}\right) \sqrt{2 x}$$ is the upper tail rate function for the random variable $(|E|- \mu)/\mu$ with speed $\sqrt{\mu}\log n$. Recall that we restrict our attention to subsets of the interval $(0, \infty)$, as our result only holds for events in which $|E|$ exceeds its expectation.
Instead of proving \[thmLDP\] directly, we will prove the following proposition instead:
\[SuperExp\] Let $$R_{t} := \left\{\frac{|E| - \mu}{\mu} \geq t \right\}\, .$$ Then $$\label{LDPInt}
\displaystyle \lim_{n\rightarrow \infty} \frac{\log \mathbb{P}[R_{t}]}{\sqrt{\mu}\log n} = -I(t) \,.$$
This statement is standard, but its proof is straightforward and we include it for completeness.
Pick $F$ to be a closed subset of $(0,\infty)$, and let $a_F$ be its leftmost endpoint. Since $I(x)$ is increasing, its infimum over $F$ occurs at $a_F$. Furthermore, $F \subset [a_F,\infty)$, and therefore, $$\mathbb{P}\left[\frac{|E| - \mu}{\mu} \in F \right] \leq \mathbb{P}\left[\frac{|E|}{\mu} \in [a_F,\infty) \right] = \mathbb{P}[R_{a_F}]\, .$$ Taking the logarithm, dividing by $\sqrt{\mu} \log n$, and applying gives the upper bound for $F$.
Next, take $G$ open and pick $b \in G$. For some $\varepsilon >0$, $[b, b +\varepsilon) \in G$. Therefore, $$\mathbb{P}\left[\frac{|E| - \mu}{\mu} \in G \right] \geq \mathbb{P}\left[\frac{|E|- \mu}{\mu} \in [b,b + \varepsilon) \right] = \mathbb{P}[R_{b}] -
\mathbb{P}[R_{b + \varepsilon}]\, .$$ Applying twice, we deduce that, for any $\delta >0$, there is an $n$ sufficiently large to ensure that $$\begin{aligned}
\mathbb{P}\left[\frac{|E| - \mu }{\mu} \in G \right] &\geq \exp \left(-(1 + \delta) \cdot I(b) \cdot \sqrt{\mu} \log n \right) \\
&\qquad - \exp \left(- (1 - \delta) \cdot I(b + \varepsilon) \cdot \sqrt{\mu} \log n \right)\, .\end{aligned}$$ Picking $\delta$ sufficiently small (as a function of $\varepsilon$) ensures that the second term is smaller than half the first term. Taking logarithms and dividing by $\sqrt{\mu} \log n$ establishes the lower bound on the probability of $\{(|E| - \mu)/\mu \in G$, and establishes \[thmLDP\].
We prove the estimate on the probability of $R_t$ in two parts. First, we show the upper bound. Fix $\varepsilon >0$. For an arbitrary pair of events $A$ and $B$, assume that, conditional on $A$, the event $B$ occurs with probability at least $1 - \varepsilon$. This implies that $$\mathbb{P}[A] \leq \left(\frac{1}{1- \varepsilon}\right)\mathbb{P}[B]\, .$$ By Theorem \[thm1\], there exists a sufficient large $n$ such that conditioning on $R_{t}$ implies that the random geometric graph has a clique of size at least $\sqrt{2 t \mu}(1 - \varepsilon)$ with probability at least $1 - \varepsilon$. This means that, for any $s$, the $s$ graded model includes a maximal clique set $\mathfrak{P} \subset T$ with at least as many vertices as the clique of the random geometric graph. Since every maximal clique set has $\tilde{\tau}_s$ indices, there can be at most $m^{d \tilde{\tau}_s}$ distinct maximal clique sets; this is an egregious overcount, but we have no need for finer control. Thus, by the union bound, the probability that there exists a maximal clique set with $\sqrt{2t\mu}$ vertices is bounded above by $m^{d
\tilde{\tau}_s}$ times the probability that a single one has the same property. The number of vertices in a maximal clique set is distributed as a Poisson random variable of mean $\tilde{\tau}_s \mathcal{D} = w$. Therefore, the chain of implication allows us to conclude that $$\mathbb{P}[R_{t}] \leq \left(\frac{m^{d \tilde{\tau}_s}}{1- \varepsilon}\right) \mathbb{P}\left[\text{Poisson}(w) > \sqrt{2 t \mu} (1 - \varepsilon)\right]\, .$$ Let $v := \sqrt{2t\mu} (1-\varepsilon)$. Applying to the right-hand side above gives $$\begin{aligned}
\mathbb{P}[R_{t}] \leq & \left(\frac{m^{d \tilde{\tau}_s}}{1- \varepsilon}\right) \exp\left( - v \left[\log\left(\frac{ v}{w}\right) -1 \right] + w
\right)\\
& \leq \exp \left( - (1 - 2 \varepsilon) \sqrt{2 t \mu} \log \left[\frac{\sqrt{\mu}}{w}\right]\right) \, ,\end{aligned}$$ where the second inequality follows for all sufficiently large $n$ by noting that all the missing terms vanish in comparison to $\sqrt{\mu} \log n$, and can therefore be absorbed at the cost of changing $\varepsilon$ to $2 \varepsilon$. By the definitions of $\mu$, $p$ and $w$, $$\frac{\sqrt{\mu}}{w} = n^{(2-p)/2} h(n) \, ,$$ for some $h(n)$ varying more slowly than any polynomial. Therefore, $$\frac{1}{\sqrt{\mu} \log n} \log \mathbb{P}[R_{t,n}] \leq - (1 - 2 \varepsilon) \left(\frac{2-p}{2}\right) \sqrt{2 t} + \frac{\log h(n)}{\log n}\,.$$ Since $\varepsilon$ is arbitrary, we conclude the desired upper bound.
For the lower bound, we replicate the argument found in the proof of \[LLowerBound\]; unfortunately, the $s$-graded model approximates the edge count of the random geometric graph from above, not from below, and so we cannot simply apply that lemma directly to deduce the lower bound. Specifically, we find a configuration which implies $R_t$. Fix $B$, a ball of diameter $r$, and let $H'$ be the event that there are $\sqrt{2 t \mu} + n^z$ vertices in $B$, where $z$ is defined as in the beginning of Section \[probanalysis\]. Since the number of vertices in $B$ is a Poisson random variable of mean $n \tau$, we can explicitly compute that $$\begin{aligned}
\mathbb{P}[H'] &= \mathbb{P}[\text{Poisson}(n \tau) > \sqrt{2 t \mu} + n^z] \\
&\geq \mathbb{P}[\text{Poisson}(n \tau) = \lceil \sqrt{2 t \mu} + n^z \rceil]\, .\end{aligned}$$ As discussed in Section \[probanalysis\], this shows that $$\mathbb{P}[H'] \geq \exp \left( - [\sqrt{2 t \mu} + 3n^z] \log \left[\frac{\sqrt{2 t \mu} + 3n^z}{n \tau}\right]\right)\, .$$ Absorbing all terms that grow in $n$ more slowly than $\sqrt{\mu} \log n$, we can deduce that, for sufficiently large $n$, $$\label{phbd}
\mathbb{P}[H'] \geq \exp( - (1 + \varepsilon)I(t) \sqrt{\mu} \log n )\, ,$$ where we bound the rate of growth of $\log [\sqrt{\mu}/(n\tau)]$ using the definitions of $p$ and $\tau$.
The event $R_{t}$ will follow if the number of edges with at most one endpoint in $B$ exceeds $\mu - 2\sqrt{2 t\mu} \cdot n^z$. Let $|E|'$ be the number of edges with no endpoints in $B$. Letting $1_{i,j}$ be the indicator of an edge between vertices $i$ and $j$, we can see that $$\begin{aligned}
\mathbb{E}(|E|') & = \mathbb{E}\left[ {N \choose 2} \mathbb{E}(1_{1,2} \cdot 1_{v_1, v_2 \not \in B} \mid N) \right]\\
&= \frac{n^2}{2} \mathbb{P}(\{\|v_1-v_2\|\le r\} \cap \{v_1, v_2 \not \in B\}) \, ,\end{aligned}$$ where $N$ is the total number of point in the torus, as before, and the probability measure in the second equality is given by the uniform process. For notational convenience, let $1^B_{i,j}$ be the indicator of the event $\{\|v_i-v_j\|\le r\} \cap \{v_i, v_j \not \in B\}$, and $\mu^B$ be its expectation under the measure of the uniform process (by symmetry, this is independent of the indices $i$ and $j$). If $v_1$ is at a distance greater than $r$ from $B$, the second condition holds trivially. For a fixed $B$, the probability that $v_1$ is within distance $r$ of $B$ is a constant multiple of $r^d$. Thus, $$\mu^B \geq (1 - Cr^d) \nu r^d\, ,$$ for some $C$ that depends only on the norm and the dimension. Thus, the expected value of $|E|'$ is bounded below by $\mu (1 - C r^d)$. Therefore, by definition of $z$, we have the inequality $$\label{expbd}
\mu - 2\sqrt{2 t\mu} \cdot n^z \leq \mathbb{E}[|E|'] - \frac{\sqrt{2t\mu} \cdot n^z}{4}$$ We now need a variance estimate for $|E|'$: $$\mathrm{Var}(|E|') = \mathbb{E}(\mathrm{Var}(|E|' \mid N)) + \mathrm{Var}(\mathbb{E}(|E|' \mid N))\, .$$ We have already calculated the expectation of $|E|'$ given $N$ above; since $\mu^B$ does not depend on $N$, we deduce that $$\mathrm{Var}(\mathbb{E}(|E|' \mid N)) = (\mu^B)^2 \mathrm{Var} \left[{N \choose 2}\right] \,.$$ A standard calculation will show that the variance of ${N \choose 2}$ is $n^3 + n^2/2$. Meanwhile, $$\mu^B \leq \mathbb{P}(\|v_1-v_2\|\le r) = \nu r^d \,.$$ Combining these facts gives $$\mathrm{Var}(\mathbb{E}(|E|' \mid N)) \leq C r^{2d} n^3 \,,$$ for some universal constant $C$.
Next, we estimate the expression $\mathbb{E}(\mathrm{Var}(|E|' \mid N))$. We can write this variance as $$\mathrm{Var}(|E|' \mid N) = \mathbb{E}\left[\Big(\displaystyle \sum_{1\le i<j\le N} 1^B_{i,j} - \mu^B \Big)^2 \mid N\right] \, .$$ We now decompose this sum into three sums by distributing the square: one sum over pairs of the form $(i,j),(k,l)$ with four distinct indices, one with pairs of the form $(i,j),(i,k)$ where one index repeats, and the final over perfect squares of terms involving $(i,j)$. The expectation of the first one is zero, as the event that $i,j$ form an edge with both endpoints outside of $B$ is completely independent of the same event occurring over distinct vertices $k,l$. For a fixed choice of $(i,j)$ and $(i,k)$, we can bound $$\mathbb{E}[1^B_{i,j} \cdot 1^B_{i,k}] \leq \mathbb{P}[ \| v_i - v_j\| \le r, \| v_i - v_k\| \le r] = (\nu r^d)^2 \, ,$$ where the first inequality follows by removing the requirement that the vertices lie outside of $B$, and thus increasing the probability. There are $N(N-1)(N-2)$ ways to choose a pair of indices that overlap in exactly one entry. Thus, $$\displaystyle \sum (1^B_{i,j} - \mu^B)(1^B_{i,k} - \mu^B) = \displaystyle \sum \left(1^B_{i,j} \cdot 1^B_{i,k} - (\mu^B)^2\right) \leq C' r^{2d} N^3 \,
,$$ for some universal constant $C'$. Again, this overestimates the value of this sum dramatically, but is sufficient for our purposes. Finally, the contribution of terms of the form $(1^B_{i,j} - \mu^B)^2$ to the sum is exactly ${N \choose 2} (\mu_B - \mu_B^2)$, which is bounded above by $C'' r^d
N^2$. Combining these results, taking expectations over $N$, and then adding the contribution of the variance of the expectation from before, we conclude that $$\mathrm{Var}(|E|') \leq C''' (r^d n^2 + r^{2d} n^3)\, ,$$ for yet another universal constant $C'''$. $r^d n^2$ grows as $n^p$, while $r^{2d} n^3$ grows as $n^{2p-1}$ - both up to factors that vanish in comparison to any polynomial. Thus, the variance of $|E|'$ is $n^p f(n)$ if $p \leq 1$, and $n^{2p-1} g(n)$ when $p > 1$, with $f(n)$ and $g(n)$ grow more slowly than any polynomial or rational function in $n$.
By Chebyshev’s inequality and , $$\begin{aligned}
\mathbb{P}[|E|' < \mu - 2\sqrt{2 t \mu} \cdot n^z] & \leq \mathbb{P}\left[|E|' < \mathbb{E}[|E|'] - \frac{\sqrt{2t\mu} \cdot n^z}{4} \right] \\
& \leq \frac{8\text{Var}(|E|')}{t \mu \cdot n^{2z}}\, .\end{aligned}$$ Regardless of the value of $p$, this quantity grows as $n^{-p/2}$ up to logarithmic factors (just as it did in Section \[probanalysis\]), and therefore, with probability $1- \varepsilon$, $|E|'$ exceeds $\mu - 2\sqrt{2 t\mu} \cdot n^z$ for all sufficiently large $n$. Following the chain of implication, we see that $$\mathbb{P}[R_{t}] \geq \mathbb{P}[H'] (1 - \varepsilon) \, .$$ Substituting the earlier bound on the probability of $H'$ completes the proof.
.2in [**Acknowledgment.**]{} The authors thank Mathew Penrose for many helpful comments. They would also like to thank the anonymous referees for their thorough reports, which greatly improved this manuscript.
[10]{}
Alon, N. and Spencer, J. H. (2008) . John Wiley and Sons, Hoboken, New Jersey.
Avin, C., and Ercal, G. (2007). On the cover time and mixing time of random geometric graphs. , 380(1), 2–22 .
Burago, Y. D. and Zalgaller, V. A. (1998). . Leningrad.
Chatterjee, S. (2010). A note about the uniform distribution on the intersection of a simplex and a sphere.
Chatterjee, S. (2012). The missing log in large deviations for triangle counts. , 40(4), 437–451.
Chatterjee, S. and Dembo, A. (2014). Nonlinear large deviations.
Demarco, B. and Kahn, J. (2012). Upper tails for triangles. , 40(4), 452–459.
Dembo, A. and Zeitouni, O. (2010). (Vol 38). Springer, New York.
Federer, H. (1996). Springer, Berlin.
Gilbert, E. N. (1961). Random Plane Networks. , 9(4),533–543.
Goel, A., Rai, S., and Krishnamachari B. (2005). Monotone properties of random geometric graphs have sharp thresholds. , 15(4), 2535–2552.
Grimmett, G. (1999). Springer, Berlin.
Hafner, R. (1972). The assymptotic distribution of clumping. , 10(4), 335–351.
Janson, S. (2004). Large deviations for sums of partly dependent random variables. , 24(3), 234–248.
Janson, S. and Ruciński, A. (2002). The infamous upper tail. , 20(3), 317–342.
Janson, S., Oleszkiewicz, K., and Rucinski, A. (2004). Upper tails for subgraph counts in random graphs. , 142, 61–-92.
Kim, J.H. and Vu, V.H. (2000). Concentration of multivariate polynomials and its applications. , 20(3), 417– 434.
Lubetzky, E. and Zhao, Y. (2014). On the variational problem for upper tails in sparse random graphs.
Meester, R. and Roy, R. (1996). (Vol. 119) Cambridge University Press, Cambridge.
Muller, T. (2006). Two-point concentration in random geometric graphs. , 28(5), 529–545.
Nagaev, S. V. (1979). Large deviations of sums of independent random variables. 7(5),745–789.
Penrose, M. D. (2002). Focusing of the scan statistic and geometric clique number. , 739–753.
Penrose, M. (2003). . (Vol. 5) Oxford University Press, Oxford.
Penrose, M. D. and Yukich, J.E. (2003). Weak laws of large numbers in geometric probability. , 13(1), 277–303.
Penrose, M. D. and Yukich, J. E. (2005). Normal approximation in geometric probability. ,5, 37–58.
Petersen, P. (2006). . (Vol. 171). Springer, New York.
Schreiber, T. and Yukich, J. E. (2005). Large deviations for functionals of spatial point processes with applications to random packing and spatial graphs. , 115(8), 1332–1356.
Talagrand, M. (1996). Concentration of measures and isoperimetric inequalities in product spaces. , 81(1),73–205.
Talagrand, M. (2003). . Springer, Berlin.
[^1]: Research partially supported by ERC AG “COMPASP" and NSF grant DMS-1005312
|
{
"pile_set_name": "ArXiv"
}
|
Cancer Therapeutics, Molecular Targets, Pharmacology and Clinical Applications--First International Conference 19-21 February 2004, Florence, Italy.
The treatment of cancer is currently undergoing a conceptual revolution due to the discovery of cancer pathways and mechanisms of oncogenesis. There are presently two complementary approaches in the field of cancer chemotherapy. Firstly, the development of "classical" anticancer agents aimed at the inhibition of cellular proliferation as a whole and bearing global cytotoxic properties that are aimed not only at cancer cells but also at all actively proliferative cells of the organism. Secondly, the development of "targeted" agents, which inhibit specific cancer targets including cyclin-dependent kinases, tyrosine kinase receptors and metalloproteases involved in cell invasion. These new molecularly targeted therapies are still experimental, with the exception of two or three compounds that have been marketed recently; a number of molecules, however, appear promising. It is generally believed that the "cytostatic" properties of classical compounds (in contrast to the cytotoxic properties) will avoid the general toxicity symptoms that usually accompany cancer chemotherapy, and that many of them will find specific indications in the cancers bearing the alteration they target. This renewal in the concept of anticancer drug discovery and development explains the emergence of new research groups involved in this field, and the continued interest in organizing meetings devoted to this rapidly expanding field. In addition, identification of the molecular targets of classical as well as novel agents, together with the completion of genome sequencing, has allowed the emergence of the science of pharmacogenomics, which tries to relate gene structure and/or expression to the activity of anticancer drugs. It is anticipated that in a few years from now cancer chemotherapy will no longer be prescribed as a function of the localization of the tumor (breast or colon, for example), but as a function of the oncogenic alteration (mutated Ras tumor, ErbB-2 amplifying tumor, etc), wherever it is localized. The paradigm of antibiotics and infectious diseases may well serve as a guide for this evolution.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Photo Illustration by Elena Scotti/Lifehacker/GMG, photos via Getty Images
So you’re ready to buy some cryptocurrency. Maybe you’ve been reading up on blockchain technology and you’re convinced it really is the future. Or maybe you watched a friend get rich off Bitcoin and you’re still kicking yourself for not doing the same.
Either way, buying Bitcoin, or one of the other many cryptocurrencies out there, can be easy and even fun. Just don’t go in with the expectation that you’ll make a million dollars overnight, and only invest money that you can live without. With that said, here’s what you need to know to get started.
Buying Cryptocurrency With Coinbase
If you’re interested in buying Bitcoin or one of the other better-known cryptocurrencies your best bet is CoinBase, which also supports Bitcoin Cash, Ether, and Litecoin. The popular digital exchange is easy to use and widely trusted, though it does go offline occasionally when trading is particularly frenzied.
You can download the Coinbase app on your phone or create an account at coinbase.com. Agree to the terms and you’ll be greeted by a chart showing the recent rise and fall of Bitcoin and other currencies. The next thing you’ll need to do is add a way to make purchases by tapping the “Buy” button on the app or clicking over to the Buy/Sell tab on the website. From here, you can connect a debit or credit card for quick small investments, or add a direct line to your bank account for larger purchases and sales that may take 4-5 days to process.
With that set, hit the Buy button again. Pick the type of cryptocurrency you’re buying and enter the amount of money you want to spend in U.S. Dollars. You’ll see how much that comes to in Bitcoin (or whatever cryptocurrency you’re buying) along with a small fee from Coinbase. Finally, confirm the transaction by hitting the Buy button and you’re done. (If this is the first time, you may get a call from your bank asking to verify the purchase before it goes through. So keep an eye out for that.)
Buying Cryptocurrency Without Coinbase
If you want to buy Ripple or one of the other up-and-coming cryptocurrencies, you’ll have to look beyond Coinbase. One popular option is Kraken, which supports Bitcoin and Ether, along with smaller currencies like Ripple and Stellar.
Using Kraken is a little more complicated than Coinbase, and you’ll want to do it through the website (there’s an iOS app but it’s not great). The first thing you need to do is setup an account here. Once it’s active, click on Account and pick Get Verified. You’ll need to provide your name, phone number and address by selecting Tier 2 at the very least. You may also need a valid government-issued photo ID and proof of residence (that’s Tier 3) to actually withdraw money from the exchange, depending on where you and your bank are located:
Next, you’ll need to deposit money into your Kraken account. To do this, go to Account, then Funding, and then Deposit. From here, follow the directions on Kraken’s website to activate a wire transfer before sending the money over from your bank account.
Now you’re ready to buy. Click on Account, then Trade, and then New Order. Then pick the cryptocurrency you want and the government-backed currency you’re using to buy it from the drop down menu in the top right corner (you can also use Bitcoin to buy smaller currencies like Ripple). Select Simple and under that click on Buy. Then enter the amount you want of whatever cryptocurrency you’re buying and hit the green button to confirm.
Ripple.com
Again, you may get a call from your bank to confirm the purchase, but that’s it. You just bought some cryptocurrency.
|
{
"pile_set_name": "OpenWebText2"
}
|
rewrite all StyleRt/*
allow url http://www.google.com/*
styleRt
|
{
"pile_set_name": "Github"
}
|
Interleukin-6 suppression of neutrophil apoptosis is neutrophil concentration dependent.
Apoptosis of polymorphonuclear leukocytes (PMNs) is a critical step in the resolution of tissue inflammation. PMN apoptosis has been studied extensively in vitro, and diverse inflammatory mediators have been shown to modulate the process. The reported effects of interleukin-6 (IL-6) on PMN apoptosis are inconsistent; however, analysis of published studies reveals at least one discriminating factor--the use of varied concentrations of PMNs in the experimental design. Consequently, we hypothesized that the in vitro effects of IL-6 on PMN apoptosis varied with the concentration of PMNs in culture. PMNs isolated from healthy human donors were cultured at concentrations from 1 to 20 x 10(6)/mL, and incubated with IL-6 doses from 1 to 100 ng/mL. PMNs cultured at 1-5 x 10(6)/mL were unaffected by IL-6; in contrast, IL-6 inhibited apoptosis in PMNs cultured at 10-20 x 10(6)/mL, compared with untreated similarly concentrated PMNs. These data suggest caution in interpreting in vitro studies of apoptosis; on the other hand, appropriately designed experiments may help elucidate the regulation of apoptosis in vivo.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
An unofficial Marine Corps birthday get-together took an odd turn in Florida Thursday night, when an anti-Trump protest walked by the bar where they'd gathered.
The Marines got into a heated shouting match with the protesters, but a strong police presence prevented things from getting out of hand.
WFTS-TV in Tampa first reported on the incident, which occurred at an Irish pub in Ybor City on Thursday.
About 100 Marines were on the first stop of an unofficial pub crawl to celebrate the Marine Corps' 241st birthday — the site of the Corps' first recruiting drive was in a tavern, after all — when a crowd of about 150 protesters passed by, said Johnny Gomes, who was in the Irish pub at the time.
"There were people with their signs yelling about fascism," Gomes told Marine Corps Times on Friday. "There were some yelling profanity. There were some yelling 'Black Lives Matter;' 'Trump is not our president.' So there was a big mix."
The Marines at the pub "didn't really care for that" and started yelling back at the protesters, Gomes said.
"Nothing happened other than a lot of screaming and hollering and the Marines getting excited and hollering out 'America,'" Gomes said.
None of the roughly 300 people involved in the confrontation were arrested or injured, police told Marine Corps Times. No fights were reported. The altercation was limited to shouting.
The whole incident lasted between 20 and 30 minutes before the protesters left, said Gomes, who credits Tampa police with preventing the volatile situation from exploding into a brawl.
× Fear of missing out? Sign up for the Marine Corps Times Daily News Roundup to receive the top Marine Corps stories every afternoon. Thanks for signing up. By giving us your email, you are opting in to the Marine Corps Times Daily News Roundup.
"I think that if the police hadn't been there, probably it could have," he said. "I'm not saying it would have but it could have. There's strong feelings on both sides."
|
{
"pile_set_name": "OpenWebText2"
}
|
The show is back after an unexpected (ahem) week off! In the first half, Brandon, Harrison and Josh discuss what they’ve been playing lately (including more Persona 4 Golden and Dust: An Elysian Tail), and discover how long it takes to finally warm up to Deep Space Nine. Then, after the break, Heidi joins us for a new segment in which some incredibly sweet ice wine is enjoyed during a session of mandatory positivity. Then, to wrap things up, we tackle a heap of listener emails and read not one, not two, but three JKP fan fics.
If you guys want to see kind of the game that Harrison was thinking of making I would say Phantom Crash (the first one on PS2) is close. Future Pop(Bubblegum crisis like) society about people who make custom mech suits that fight in arenas and you AI are hyper intelligent animals (cats, dogs, rabbits, anything) It’s an awesome game mainly because it removes the clunky action from a game like mechwarrior and replaces it with something like the gameplay from the Ghost in the Shell game (the tachikoma one) from the ps1 era. I wish that game could be made today because I spent hours tooling with all the things you could do.
|
{
"pile_set_name": "Pile-CC"
}
|
Foul play is not suspected in the death.
There has been shock and sadness at the sudden death of a PSNI officer at a Co Antrim police station.
He was Inspector Peter Magowan and he died in an incident at Ballymoney police station on Monday.
The man's colleagues were informed on Tuesday.
Superintendent Tony Callaghan added: "Police are investigating the sudden death of a 55-year-old police officer in Ballymoney Police Station on April 18.
"We are not treating the death as suspicious.
"The Police Ombudsman office has been informed."
North Antrim DUP Assembly candidate Mervyn Storey said: "As someone who lives in Ballymoney, there is real shock in the town that a police officer has been found dead in the local station.
"My team actually had a meeting in the station yesterday to deal with some local issues. Whilst I didn’t know the officer personally, I knew of him. My prayers are with his family and the wider police family in Ballymoney.”
North Antrim MP Ian Paisley described it as "very sad occurrence".
So sorry to hear of the death of a police officer in Ballymoney. Thoughts and prayers are with the officer's family at this time — David McIlveen (@DavidRMcIlveen) April 19, 2016
Belfast Telegraph
|
{
"pile_set_name": "OpenWebText2"
}
|
Is there a place for project management in Agile?
There is debate on the role of management in Agile and Lean projects - is there a place for project management in Agile? Well indeed there is because agile methodologies are in essence nothing other than a set of management frameworks.
What makes a good agile team/’s manager, development manager, chief architect, programme manager or whatever title they are given. The titles are a bit misleading so let’s not start not with a title. Instead let's use a job description to make it easier to decide where this manager fits into an agile framework. Putting the title to one side, the role we are looking at, is the one held by the person who is responsible for leading the team (or teams) to a product release.
They are the link to the customer, or for internal teams the link to the business. They are the ones who are pivotal in translating a business need into a product that delivers value. To them the team is everything for without it there is no product. This is not to say they do not serve the business as well. They serve both by communicating the needs of the business to their teams.
Below I will refer to them as development managers.
Negotiating
So now we know who they are let’s look at the tasks they have to perform. The main one is arbitration, they need to act as the umpire. Smoothing conflict starts by negotiating a path between the desires of the business and the teams who will deliver them. They need to be adept at recognising what desires will deliver value and which ones will not. Above all they work with people.
The development manager will be responsible for requirements elicitation, communication, delegation, facilitation, recognising and managing risk, team building and recruiting. They strive to protect their teams from feature creep, bad smells, anti-patterns. They make opportunities for code reviews, they also make opportunities to obtain frequent feedback.
When their teams have an internal difference it is the development manager who can help them choose the right option. To do this they use their advantages of:
visibility of the wider picture
experience of similar decision points
technical knowledge
understanding of the dynamics of the team
They use all of these attributes to moderate the debate and assist the team to find the right solution.
The development manager needs to negotiate the scope of the work the resources required and the time frame to delivery. It is best to keep the scope to the minimum that is viable for value to be gained. Three fast and small but good releases are much better than one large one. The customer may want an all singing all dancing release and this will bring a need for negotiation.
I was involved in a start up business within the group of companies I worked for. The newly installed MD wanted an application with all the features and then some. Feedback was sought as functionality was delivered. He became insistent that without all features he would not accept anything.
He also chopped and changed what this set of features were. This was mainly a ruse to cover the fact that his startup had no traction and was not delivering any sales. While the team welcomed change they were getting frustrated by the apparently arbitrary shift from one thing to another. The negotiation winner here, was to deliver a small product that drove sales.
Development managers lead, they do not write code.
Some hold the view that a development manager should also write code, this is bull. They need to understand code, they need to read code, they will probably have written code but that should remain in the past. They may be a test engineer or a another type application expert.
They need to be technologically knowledgeable smart, aware, experimental and enquiring. They need to continuously improve their knowledge on the subject. They must like the users of the systems they are involved in building. They will have belief and be evangelical of agile and lean principles.
There are very good programmers who are not very good at communicating, the classic introvert. There are very good programmers who are extrovert and just love gassing away. There are those who may have been good programmers but have transferred to a leadership role. The development manager does not need to distracted by writing code good or bad. This is an activity they need to delegate.
Let the development manager leave writing code alone and concentrate on leading those who are actively programing. They will on occasion dip back into getting their hands dirty but they will do it as a scientist not as an engineer. The scientist experiments to find answers the engineer builds solutions.
The development managers interest in new technology may lead them to undertake empirical investigation. For instance they may pick a technology up, take it to hello world and beyond. In doing so they will evaluate the usefulness of different emerging technologies and trends. If these offers solutions they are in a position to introduce it to their teams.
The rules are there are no rules
There are rules that applied to the leadership however. Sometimes we referred to them as the non rules. These non rules were really just guarantees as to what the teams were entitled to expect. There were only a handful and all of them had the number 10 in them.
Developers are bad at looking after their own self development or seeking help. One non rule we had was the 10% rule. This was where 10% of the working week (3.5 hrs) is made available for self development. This half day study time was slotted in for learning a new skill or honing an existing one. Individuals pairs or small groups had time allocated for discovering new technology or soft skills.
The 10% rule included explorations into new hardware and processes as well as exploring new languages or trying out a new types of data stores. I have had to drag developers kicking and screaming into utilising this time. The development manager will play a valuable part in the quest to amplify learning.
The development manager will act as the enabler for self learning processes. They will take suggestions on what the team would like to look into further. As a scientist not a developer they may vet new developments in code or databases, they will benchmark and assess. They will prepare the ground for further experimentation and at the same time be pragmatic to avoid trendiness.
The development manager will set up the environments to run whatever the learning process requires. They will establish playground or find an online playpen. They will do a bit of research and sift through what background or resources are available to find the most useful sources. They may take it to ‘hello world’ and then hand over a curated environment to team members to evaluate and play with.
Another non rule was the 10 minute rule. if you are stuck on a problem for longer than 10 minutes ask someone for help. Articulating the problem would most often fix it. The first instinct of a development manager writing code will be to grab the keyboard and try fix the problem. Worse still they may take it on themselves to fix, deeming it too difficult for the skills of person who asked for help. Big mistake, people learn nothing from seeing someone else do it. They learn from discovering how to do it themselves.
Respect
The development manager is not hands on, they are there to help. Leading a colleague to a solution is helpful. Encouraging them to see through the problem and to find a solution for themselves is even more helpful. If the solution is beyond the skill and knowledge of the development manager all well and good. They can research, learn from the team or simply note that this team member has the answers up their sleeve.
The team will respect the manager for not talking down to them and for enabling them, for supporting them and for encouraging them. If they have to write production code they will not have the focus or the time to give enough to their project team. Their project teams will need them focused on the team’s needs and on the needs of the business.
It is said that the team will have more respect for someone also writes code. My experience shows this not to be the case. The last thing the team needs is a part time coder nostalgic to get their hands dirty. They will not respect the coder who is their on the presumption they are a better coder than themselves.
If the development manager tries to multitask the they will resent the time they are not writing code. They will be distracted, they are not delegating. Their teams will resent not having the support they need. If their background was testing or infrastructure engineering or application training there is possibly less temptation for them to get their hands dirty. Yet the focus must be on the team and the product not their history.
The reasons development managers manage is to best serve their teams. It is not to establish a management centric culture. It is not to build hierarchies it is to help, assist and make their team’s lives easier. Management is providing opportunities for their teams to excel, improve and learn. To deliver value to the customer. It is acting as a link between the stakeholders and the production teams.
Style
I have been responsible for quite a few development managers and they each have their own style yet they all share some stylistic attributes. They all demonstrate clearly and openly the trust they have in their teams and the individuals within them. They are comfortable in one to one discussions with team members and in group discussions. They all have two ears and one mouth. That is they listen more than they talk.
They also appreciate the style of their teams and its members. The recognise the learning styles of the individuals in the team and are then able to adjust how they deal with the individuals by adjusting their own style. The can also use this knowledge to adjust the team to make it rounded and get a good mix of personalities. With insight into style the team dynamics can be optimised.
Flavours
Style also covers the type of agile framework that is best for the team. Some teams will naturally go for one or other methodology. Some will craft their own. It does not matter too much if all teams follow the same methodology. It demands even more agility on the behalf of the development manager if they all use different frameworks. I would advise against imposing a framework, let team find one.
The team should decide what style of working suits them the most. XP, FDD, Scrum, Kanban, Lean etc. all have agile principles. These main frameworks have enough core tenets on which to build an agile way of working. Agile methodologies need to be agile in themselves, I find Scrum the least agile of all the ones listed above yet it is popular.
Rules v’s culture
All agile principles propose that teams should be self organising and that they should be empowered so let them choose, develop or mutate their own style. In other words a style should fit around the team. The team should not be fitted into the style.
The need for established boundaries will still exist. It is important to remember this is not the wild west. There are boundaries, anarchy is not to be encouraged, anti patterns are taboo, risk is to be avoided or mitigated, best practice must prevail.
The best way of ensuring boundaries are drawn within the territory of best practice and excellence is to develop, instill and then continue to influence the culture. A development manager will; add to and play a big part in forming the culture that their teams operate within. Cultures need to be built, modified, reinforced and nourished. A culture will contain processes, techniques and patterns.
The culture is what transmits a vision of desirable behavior when a system of dictats and rules are not appropriate. A culture where the desire to, empower the team and trust motivated people to do their jobs, is a good replacement for rules. A self organising team establish and introduce best practice. A culture that encourages continual incremental improvement builds integrity into the product.
The structure of communications
Regardless of which agile framework is used the development manager will need to ensure that they have some core processes available and followed. Regular communication being perhaps the most important of these. The XP standup can be held sitting down as long as is not around a table and it sticks to a time limit. The idea is to keep the meeting short.
I found splitting the traditional daily meeting into two daily meetings quite useful. I named these meetings morning prayers and evensong. Daily meetings can be done remotely using any conferencing app that has text and voice or better still text, voice and video.
Daily meetings
Morning prayers were as the name suggests in a morning after everyone had settled in. It was allocated set duration of around 10 minutes (a minimum of 5 and a maximum of 15). It was used to discuss what each member of the team was going to achieve that day and what resources or dependencies they needed, e.g. a test instance on the server, dependencies, a clone of a database or clarification on how a feature should work.
Problems discovered in the stand are then taken off line to be resolved. The purpose of the meeting is to bring issues to the attention of the development manager and identify blockers. It is not desirable to try and fix the issue in the meeting. Once identified the development manager will allocate the resources to remove the problem.
Evensong was an hour before the end of play, the same 10 mins allowed team members would cover what they had actually achieved. By this point it the day everyone has a good idea of what progress they have made. It also everyone a chance to reflect on what they have done. It did this while there is 45 mins left to maybe, check over work done, remove some technical debt, clean it up and better still share it.
If the level of work completed was not what was expected to have achieved it gives a chance to point out what the blockers are. It gives the team and the development manager an opportunity to plan the work for the following day. It provides information for the planning process, updating release schedules and the road map.
Periodic meetings
On Friday afternoon there is an additional ceremony, a sort of special mass, and it has a party feel to it. Wine and bread are imminent at the review theatre. The Friday review allows everyone to put on the finery and parade their work to each other, to other teams, to the customer and to the business. It is the place from where feedback emanates. It is pat on the back time.
Friday demos might not be actually held on a Friday. They could take place anytime when the appropriate stakeholders are available. For the development manager it is a time for pride and praise all round (hopefully). For the team it is a chance to view integrations and deployments in situ.
Retrospectives
I read somewhere that retrospectives should be at the end of an iteration. If that were the case the iterations are too long. Continuous deployment encourages short iterations. Sometimes these are daily and it is possible to have multiple iterations completed in a single day. There is no need to have retrospectives daily. A retrospective is an opportunity to change the way we work and this is not a daily task.
Some functionality may not lend itself to such short iterations and the length of iterations needs to be kept to a minimum. If long iterations become frequent it is time to look at other ways of building a product. The architecture could be wrong and a design that increased the number of services may be a better solution. Ever smaller iterations are not the end of reflection or retrospectives. That dear old mantra:
What have we done well
What do done badly
What should we stop doing
What should we start doing
is a set of questions that still need to be asked periodically and can really be asked on any occasion. This reflection is the root of the retrospective process and may be done in at some depth. It can take place alongside or within a code review. Natural time periods between retrospectives will appear, as will natural break points in production.
Location
It is important to have the right location, the team must be away from their desks and not on a computer. There were times I gave the development managers my office to hold these meetings. Too many teams and my having no office put paid to that.
The location needs to have a whiteboard, it is good if it has a large monitor. I incorporated informal areas where these meetings could be held. They were also where we sometimes met with stakeholders. They were not formal conference rooms nor were they held sitting around a table.
…
Originally published on September 16 2014 It was based on notes published on an internal company blog and relates to the development managers at that company.
With recent events reminding me that once again British cycling is a beacon for teamwork and continual incremental improvement. Derived from Dave Brailsford's framework of the "aggregation of marginal gains" where individuals are charged with breaking down and improving individual elements of their technique and ability seeking to improve their overall performance, the British team taking 12 medals at Rio is as good a place as any revisit strategies for success.
It is not intended as a formula or a prescription to follow. It is an simply an example of how this management role has worked in the past. I was lucky to have a couple of star managers grow into this role, one came from a technical product training background the other from a testing background. Both knew enough about code to understand it. Both were astute technologists, both embraced continuous improvement and continuous integration. Yet the reason they were both stars was because they were very people focused. They also liked users.
|
{
"pile_set_name": "OpenWebText2"
}
|
But in the Trump administration, any ridiculous claim emanating from the Oval Office must be translated into official action, and now that we have an attorney general eager to indulge and validate the president’s most corrupt impulses, that’s exactly what’s happening. The Post reports:
Attorney General William P. Barr has tapped John H. Durham, the U.S. attorney for the District of Connecticut, to investigate the origins of the special counsel’s probe into Russian interference in the 2016 election. Barr picked Durham in recent weeks to work on the review, which is designed to ensure the U.S. government’s “intelligence collection activities” related to the Trump campaign were “lawful and appropriate,” a person familiar with the decision said.
What’s so absurd about this is that we know exactly what the origins of the investigation were. It began when Trump campaign adviser George Papadopoulos (one of multiple Trump associates who would later plead guilty to crimes) bragged to an Australian diplomat during a night of drinking that Russia had damaging information on Hillary Clinton that was being passed to the Trump campaign, and Australian officials then notified the FBI.
AD
AD
By all accounts, the investigators proceeded with unusual care, deeply concerned about what Russia was doing, but just as deeply concerned that their investigation might affect the outcome of the election if its existence became public. Which is why they successfully kept it secret until after the election was over, a fact that is utterly incompatible with the Republicans’ theory that there was a sinister conspiracy within the Justice Department that was bent on destroying Donald Trump.
What quickly became apparent in that investigation was that not only was Russia mounting a comprehensive effort to help Trump get elected, there was a very strong possibility that the Trump campaign was cooperating with the Russians in some form. During the campaign, there were anywhere between 100 and 250 contacts between Trump campaign figures and people associated with Russia. The Trump officials making those contacts included his son, his son-in-law, his campaign chairman, his personal lawyer, his future national security adviser, and his future attorney general, among others.
We can argue about how you might define “collusion” and whether what happened constituted a conspiracy. But from the standpoint of counterintelligence investigators looking for a fire, there was smoke billowing from every window in Trump Tower.
AD
AD
Which brings us to something we have to stop and emphasize, because it’s so illogical and even bizarre that your brain might reject it outright — despite the fact that it’s staring you in the face.
The Republican position on the Russia scandal is that there should never have been an investigation in the first place. This seems almost too insane to believe, but it’s true.
Republicans actually want to claim that when the FBI first got hints of what was going on, their response should have been not “This is something we should check out,” but rather, “Eh, that doesn’t seem like any big deal. No need to look into it.”
This is still their position, even given everything we’ve learned since — all the details about the Kremlin’s effort to help Trump; the eagerness of Trump, his family, and his campaign to get help from Russia; and all the convictions and guilty pleas of Trump associates.
AD
AD
So now, even after the copious misdeeds documented in the Mueller report, Republicans still cling to the idea that there should never have been any investigation, and the very fact that the investigation began is evidence of malfeasance in the FBI.
For the record, I have no reason to believe that Durham is a partisan hack, as is the attorney general who tasked him to lead this new inquiry. But I can offer a prediction about how this is going to go.
Durham will find that the investigation was completely justified and carried out in accordance with counterintelligence practice. He may, however, find a few instances of minor sloppiness as you would in any investigation, with some agent somewhere forgetting to fill out his 27B/6 form in triplicate.
AD
Barr will then repeat the strategy that worked so effectively with the Mueller report, holding a news conference before anyone has seen the actual results, during which he will say that serious misconduct was uncovered. Trump will then triumphantly proclaim that it has been proven that he was the victim of a Fake News Hoax Witch Hunt, a claim that will be echoed loudly on Fox News and conservative talk radio.
AD
In the end, this investigation of the investigation will be added to the pile of documents and conspiracy theories that Trump and his defenders believe prove his innocence but do nothing of the sort (perhaps they can file it next to the [Devin] Nunes memo). But what should really disturb us is the fact that the Justice Department has basically become a servant of the most corrupt and dishonest president in living memory. Who knows what he’s going to order it to do next?
|
{
"pile_set_name": "OpenWebText2"
}
|
Therapeutic options in the management of incidental carcinoma of the prostate.
Incidental carcinoma of the prostate is a protean disease with a natural course which may be indolent or aggressive, with prognosis correlated with histologic grade and extent of disease. Treatment of this pathologic entity has varied over time and has been governed by institutional policy rather than randomized comparison of therapies. This report reviews the literature on incidental prostate cancer focusing on outcomes of patients as related to different therapeutic maneuvers. Observation alone with careful follow-up is appropriate therapy only for those patients with well differentiated disease of limited extent. Patients with diffuse or less differentiated disease required definitive therapy to prevent symptomatic progression. Hormonal manipulation alone has not been demonstrated to be of benefit. Radioactive implants have yielded poor disease-free survival. Radical prostatectomy by an experienced surgeon for patients with adequate health to tolerate the procedure has been associated with acceptable morbidity and excellent local control and survival. Radiation therapy has yielded similar excellent local control and survival and appears to be appropriate for a broader range of patients regardless of health or age.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Illustration de Marseille en 1720 pendant l'épidémie de peste. — MARY EVANS/SIPA
Deux archives datées de 1720 et 1721, exhumés par des généalogistes, témoignent de la manière dont la France a géré la dernière grande épidémie de peste.
L’historien Patrice Bourdelais, spécialiste des épidémies, explique les différences et les ressemblances des méthodes, à 300 ans d’écart.
Ce sont deux documents d’archives qui datent de 300 ans. Et pourtant, ils font écho de façon surprenante à la situation de confinement et de lutte contre le coronavirus que vit actuellement la France.
Le premier est un « passeport de santé » daté de 1720 (voir la photo ci-dessous), en pleine période de peste. Le second, édité un an plus tard en 1721, rassemble les instructions générales envoyées depuis Paris pour juguler la propagation de ce fléau dans les « lieux attaqués de la maladie contagieuse ». A trois siècles d’écart, on s’aperçoit que certaines méthodes ont évolué, mais d’autres non. Interrogé par 20 Minutes, l’historien des épidémies, Patrice Bourdelais, nous aide à y voir plus clair.
Exemple de «passeport de santé» daté de 1720 et qui permettait à l'intéressé de voyager. - PicMonkey / 20 Minutes
La peste de 1720-1722, dite « de Marseille »
En mai 1720, un bateau de commerce transportant des étoffes de coton du Moyen-Orient accoste à Marseille. Longtemps, ce navire a été accusé d’avoir importé la peste dans le sud de la France. En 2016, des études génétiques sur le bacille infectieux ont émis une autre hypothèse : une résurgence de la «peste noire» qui avait décimé la moitié de l’Europe au XIVe siècle.
Quoi qu’il en soit, cette épidémie a duré plus de deux ans. Elle a tué près de la moitié des 85.000 habitants de Marseille et fait entre 90.000 et 120.000 morts en Provence et Languedoc, selon les estimations. « Il s’agit de la dernière grande peste d’Europe occidentale, souligne Patrice Bourdelais. A l’époque, la population avait une grande expérience de ces épisodes épidémiques qui survenaient tous les 5 à 10 ans. L’alerte était forte. Ce qui n’est plus le cas aujourd’hui. »
L’efficacité du cordon sanitaire
Exhumée des archives de la bibliothèque nationale de France (BNF) par un groupe de généalogistes, une note d'« instructions sur les précautions » à prendre, datée de 1721, expose notamment la manière dont un cordon sanitaire militarisé doit être organisé pour faire blocus autour d’une zone contaminée.
« Le système était très bien rodé avec un cordon sanitaire pouvant rapidement reculer ses lignes de contrôles au fur et à mesure de la progression de la maladie, explique Patrice Bourdelais. On a pu isoler efficacement une partie du territoire en 1720, c’est une grande différence avec aujourd’hui ».
Et l’historien de préciser le dispositif : « Rigueur du blocage grâce à la vigilance sans faille des soldats, organisation de postes afin d’alimenter le lieu, rigueur dans la chasse aux repaires possibles de l’épidémie comme les paillasses, les vêtements ou les étoffes. Car la peste se développe d’abord dans une population de petits rongeurs qui transmettent la maladie à la puce qui pique l’homme et le contamine. »
Ressemblances frappantes
Pour l’historien, la similitude la plus frappante entre les méthodes d’aujourd’hui et celles d’il y a 300 ans, ce sont surtout les trois niveaux de distinction des malades. « L’organisation du placement des malades en trois strates est clairement exprimée dans ce document : les infirmeries pour les malades, des lieux pour les convalescents et d’autres endroits différents – où l’on effectue la quarantaine classique – pour ceux qui ont été en contact avec des malades et sont donc suspects. Cela rappelle des débats actuels sur le confinement général où nous sommes tous devenus suspects. »
Autre ressemblance, le « passeport de santé » évoqué aujourd’hui pour accélérer le déconfinement. En 1720, il s’agissait d’un laissez-passer, différent de l'« autorisation de déplacement dérogatoire » de 2020. « C’est une sorte de certificat de dépistage collectif. Les habitants d’une commune qui n’était pas touchée par l’épidémie pouvaient obtenir une autorisation de circuler. »
Le document montre l’exemple d’un Alexandre, de taille « médiocre » et aux cheveux « châtains », vivant à Remoulins dans le Gard. « La peste sévissant à Avignon, non loin de là, ce qui rendait les habitants du secteur a priori suspects. Le magistrat devait donc délivrer une autorisation pour que la personne puisse se rendre dans un territoire qui n’était pas touché. Visiblement le Tarn, dans ce cas. »
Autre temps, autres mœurs
En revanche, certaines instructions de l’époque peuvent choquer. Illustration avec deux passages du document : « Si quelqu’un de ceux qui ont été consignés vient à s’échapper, ils lui feront casser la tête » ou encore « On doit faire tuer tous les chiens et tous les chats, tant au dedans qu’au dehors du blocus, à une lieue au moins (…). Quoique ces animaux ne prennent pas le mal, ils le communiquent très souvent ».
« La rigueur du traitement réservé à ceux qui essaieraient de traverser les cordons, ainsi qu’à nos animaux domestiques, s’explique par la volonté absolue de limiter les déplacements, et l’observation répétée que ces animaux semblent aussi transmettre la peste, car leurs puces sont responsables de leur malheur », observe Patrice Bourdelais.
L’aide aux démunis
Mais la solidarité existait déjà à l’époque. Ainsi, les « instructions générales » se terminent par ce paragraphe : « Les amendes qu’il faut ordonner souvent pour les contraventions qui ne méritent pas la mort seront appliquées aux pauvres du lieu. Et le meilleur usage qu’on en puisse faire est de les employer à leur acheter des habits, au lieu de ceux qui auront été brûlés ».
« Effectivement, tous ceux qui ont dû laisser brûler leurs vêtements, voire leur maison, sont secourus par la communauté, la charité et les aumônes. Les œuvres chrétiennes étaient très nombreuses, rappelle l’historien. Et le montant des amendes infligées pour avoir contrevenu à certaines règles peut aussi être mobilisé à cette fin. » Aujourd’hui, on appellerait ça une subvention ou une aide de l’Etat.
|
{
"pile_set_name": "OpenWebText2"
}
|
Q:
I am trying to map IDs from inside a list of objects into a new list using React js
I want to make a new list from an existing one by fetching the IDs from the existing list of objects and copy it into a new list which is accessible throughout the component.
This existing list is being returned as part of response of a BE API
const key = [];
await RequestPortInService.fetchCurrentServices().then(response => {
response.currentServiceList.map(key => (
key = response.currentServiceList.serviceId
));
alert(key);
})
A:
your key values in the object response.currentServiceList.map will be on key so try using
var newSerivceIds = {};
await RequestPortInService.fetchCurrentServices().then(response => {
newSerivceIds= response.currentServiceList.map(key =>
{return key.serviceId});
})
alert(newSerivceIds)
|
{
"pile_set_name": "StackExchange"
}
|
A furniture store is giving Moncton-area residents the chance to recycle K-Cups and help create jobs for people with disabilities in the Maritimes.
Wheaton's furniture has launched a pilot project to recycle the convenient coffee cups and at the same time, create jobs for people who are disabled.
The single-use cups are popular, but the handy coffee pods generate a significant amount of waste.
Garnet Wheaton, the founder of Wheaton's, said he believes that looking for a way to recycle the cups is the responsible thing to do.
We know that K-Cups produce a lot of waste to the landfill, so we've decided to take some action. - Garnet Wheaton, store founder
"We want people to be able to enjoy coffee and enjoy the conversation that goes with coffee. And we want them to be able to do that in an environmentally and responsible way," he said.
"We know that K-Cups produce a lot of waste to the landfill, so we've decided to take some action."
The Wheaton's store in Moncton is accepting K-Cups from any source for recycling. Anyone bringing in 24 used coffee pods will receive a $1 coupon for Mother Parkers coffee products.
The pilot project is taking advantage of a gap in the store's transportation system.
Trucks delivering furniture to Moncton area normally return to Halifax empty. But now, those trucks will haul the collected empty cups back to Halifax.
"We had the transportation system in place to be able to collect and to deliver these cups back to recycling without burning extra fuel," said Wheaton.
"It's all done in a very efficient manner."
Workers at the Dartmouth Adult Services Centre remove the coffee grinds from used K-Cups so they can be used in compost or fertilizer while the plastic cups are recycled. (CBC) In Halifax, the empty coffee packages are turned over to the Dartmouth Adult Services Centre, where employees break down the cups.
Cathy Deagle-Gammon, the centre's executive director, says it's great employment for adults with intellectual disabilities.
"It's a project that helps our clients to gain skills," she said.
"It's variety in the work that we do and I think the other thing to everybody likes to help contributing to bettering our environment."
The leftover coffee grinds make their way to compost and fertilizer while the plastic cups can be recycled.
"There is absolutely nothing that isn't great about this project," said Deagle-Gammon.
If the pilot project catches on, Wheaton said he hopes to expand the program to all of the company's stores in the Maritimes.
|
{
"pile_set_name": "OpenWebText2"
}
|
Jamaica's Rum
by Denise Salmon
(Jamaica)
Jamaica's Rum
Jamaica's Rum is said to be the best in the world.
It is certainly the more 'spirited' part of Jamaica's history :-)
Rum is made from spirits obtained from distilling the fermented residue of the sugar cane. The process requires lots of patience and time, as it requires fermentation, distillation and the right amount of aging.
The tasty liquid was discovered when workers crushed the cane and found out that the fermented residue was sweet and quite stimulating.
They first called the liquid liquor of cane before it was named Saccharum, the now scientific name for sugar cane.
It quickly became famous as it was loved by seafaring men, who carried it across the Atlantic to Europe where the Upper class British folks called it "Kill Devil" - and thought of it as a drink for the lower class people.
But in the 17th Century it was named "Rumbullion" an old English term for uproar or rumpus. And then it was shortened and called rum.
In 1909 however England's Commission on Whiskeysaid Jamaica's Rum which was made with great pridewas a spirit and then Jamaica was able to exporta little over a million gallons of the rum!
But the process through which rum is made is supervised as it is aged in wooden barrels whichadds color and flavor.
It is said that Jamaica was the first island to produce rum commercially and this rum is considered to be the best in the world.
There is currently an Appleton Estate Rum tour that will provide you an good history of this aspect of Jamaica's history and also show you exactly how we process our world-admired Rum.
During the tour visitors are also allowed to sample the many different rums.
Articles You Might Have Missed...
Thanks For Reading!Like What You Just Read?
Please join me in my effort to share the beauty and uniqueness of Jamaica with the world. You can do so by sharing this insightful narrative on the social media of your choice and ask others to do the same (BELOW).
Need Help?
We have a cherry list of top Jamaican companies that might be able to help you. Please click here to see them, our preferred partners and site sponsors.
Get Exclusive Updates & Tips!
You are also welcome to join my special friends list and receive exclusive updates (like this), tips, trivia and stories from lovers of Jamaica!
|
{
"pile_set_name": "Pile-CC"
}
|
Springfield Manor Winery Wedding in Maryland | Paige & Jon
Paige and Jon are two of the sweetest people we’ve had the pleasure to work with. We first met them in Old City at their engagement shoot and you know you hit it off when we all can’t stop chatting and have to force ourselves to so we can start taking some photos! They are both just such kind, lovely human beings that working with them was nothing less than amazing.
To top it off, their Springfield Manor Winery wedding was SO perfectly autumnal and every little detail was spectacular. It was even more amazing because the leaves changed color right before their wedding day and the result was an insanely gorgeous backdrop for the entire day. Lucky us! Definitely check out some of our favorites below (it was so so hard to narrow it down)!
|
{
"pile_set_name": "Pile-CC"
}
|
Overexpression of LAPTM4B-35 is a negative prognostic factor in head and neck squamous cell carcinoma.
Overexpression of LAPTM4B-35 (lysosomal-associated transmembrane protein 4β-35) is associated with a poor prognosis in numerous malignant tumours. Expression patterns and effects of LAPTM4B-35 on head and neck squamous cell carcinomas (HNSCC) are unknown. The aim of this study was to investigate the prognostic relevance of LAPTM4B-35 in HNSCC. Tissue microarrays were constructed with primary tumours and associated lymph node metastases isolated from 127 patients. The expression of LAPTM4B-35 was investigated by immunohistochemistry and the results were correlated with survival data. LAPTM4B-35 in the primary tumour was highly expressed in 47.2% of the patients (60/127). LAPTM4B-35 expression was significantly associated with tumour stage. Moreover, overexpression of LAPTM4B-35 correlated with a significantly worse disease-free survival (10.23 years vs. not reached) and a higher recurrence rate (40.7% vs. 25%). High expression of LAPTM4B-35 in lymph node metastasis was found in 29.2% of cases. In 19.4% of cases, high LAPTM4B-35 expression was observed in both the primary tumour and corresponding lymph node metastases. In conclusion, our data indicates that overexpression of LAPTM4B-35 is associated with poor prognosis and may therefore serve as a new prognostic marker in HNSCC.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
The epithelial barrier presents a significant obstacle to the delivery of macromolecules in the size range of 20 - 150 kDa. In particular, the tight junctional complex, which links adjacent cells and occludes the paracellular space, presents a significant obstacle to delivery of macromolecules. To improve the transport of macromolecular biologics across epithelia, new approaches need to be developed that enhance paracellular drug transport by specifically and reversibly modulating tight junctions. In this proposal, we investigate the effect of nanostructured surfaces on the modulation of tight junction permeability and transport of key therapeutic molecules in vitro. We seek to determine the mechanisms through which epithelial permeability is enhanced by nanotopography and optimize nanostructured materials to broaden the types of drugs that can be delivered paracellularly. It is expected that the fundamental knowledge gained in these studies will enhance the development of new epithelial drug delivery systems.
|
{
"pile_set_name": "NIH ExPorter"
}
|
llest value in 0.03, -0.5, -1, -9, -2/95, 1?
-0.5
Which is the second biggest value? (a) 0.5 (b) 62/375 (c) -2
b
Which is the fourth biggest value? (a) 0 (b) 3 (c) 0.2 (d) 2 (e) -2/25 (f) 8
c
Which is the second biggest value? (a) 1 (b) -69 (c) -0.06 (d) -0.1 (e) -10 (f) -1/3
c
What is the second smallest value in -4, -0.9, 0.2, -7, -1?
-4
What is the fourth biggest value in 1, -2, 0.425, -1, -5?
-2
What is the second smallest value in -1/2, -0.4, 2/1079, -0.06, 0?
-0.4
Which is the second smallest value? (a) 1/2 (b) 279 (c) -0.2 (d) 4
a
Which is the biggest value? (a) 2 (b) 2582 (c) -34
b
What is the sixth biggest value in 4, -4, 66, -0.053, 0.2, 0.4?
-4
Which is the smallest value? (a) -0.5 (b) -0.31 (c) -168 (d) -5
c
What is the fourth biggest value in 3, 5/3, 0.4, 308, 5?
5/3
Which is the second biggest value? (a) 0.199 (b) -0.4 (c) -5 (d) -1/2 (e) 1 (f) 4
e
Which is the smallest value? (a) -77 (b) -0.5 (c) -4 (d) 0.4 (e) 2
a
Which is the second smallest value? (a) -0.8 (b) 2 (c) 0.2 (d) 2072
c
Which is the smallest value? (a) -16/5 (b) -2 (c) 5/4 (d) -9
d
What is the second smallest value in 3, -2/11, -194, -0.1, 0.3?
-2/11
Which is the biggest value? (a) -4 (b) 144 (c) 0.5 (d) 8
b
What is the fifth biggest value in 1/1296, 3, 29, -2/7, -3?
-3
Which is the biggest value? (a) 0.4 (b) -0.4 (c) -0.5 (d) 20/3 (e) 0.05 (f) 0.03
d
Which is the second smallest value? (a) -0.11 (b) -608 (c) -2/17
c
What is the smallest value in -50/9, 1/70, 1/13?
-50/9
Which is the second biggest value? (a) 0.14 (b) 51 (c) 1/9
a
What is the smallest value in -0.7, 124, 1/188, -3/5, -5?
-5
What is the sixth smallest value in 1/3, -0.4, 13, -2/13, -6, 27?
27
Which is the second biggest value? (a) -2/9 (b) -7 (c) -11/7 (d) -6 (e) -1/5
a
Which is the second biggest value? (a) 0.5 (b) -0.4 (c) 0.03 (d) -13 (e) -0.24
c
What is the fifth biggest value in 2/11, 2/3, 733, -5, 57?
-5
Which is the smallest value? (a) 4 (b) -0.2 (c) -600
c
Which is the second biggest value? (a) -820 (b) 0.3 (c) -1.8 (d) 5 (e) -0.4
b
Which is the fourth smallest value? (a) -6 (b) -1/3 (c) -9.32 (d) -0.5 (e) 1
b
Which is the third biggest value? (a) -4 (b) 3 (c) 2 (d) -1207 (e) -1
e
Which is the smallest value? (a) -2 (b) -14454 (c) 2/13
b
What is the biggest value in 1, 3, 2/13, -3, 2424?
2424
Which is the fourth smallest value? (a) -7 (b) -11 (c) -2117 (d) 4
d
Which is the fourth smallest value? (a) -1557 (b) 4 (c) 0.4 (d) -2/5
b
What is the biggest value in 10498, 4/5, -0.1, -4/5, -3?
10498
Which is the fifth biggest value? (a) 4 (b) -1.2 (c) -0.5 (d) -4 (e) 41
d
Which is the smallest value? (a) -1 (b) -1/9 (c) -1/49615 (d) 2
a
What is the fourth biggest value in -2/27, 1, 2, 0.05, 26?
0.05
Which is the second smallest value? (a) -0.2 (b) 30 (c) 1.3 (d) 5 (e) -7 (f) 1
a
Which is the fifth smallest value? (a) -2 (b) 0.1 (c) 1/3 (d) -312/29 (e) 0.5 (f) -3/2
c
What is the third biggest value in 0, -23, -1/97, -2/3?
-2/3
Which is the second biggest value? (a) -2 (b) 3.5 (c) 12.42 (d) -5
b
What is the third biggest value in -0.3, 13147, 4/7, 0.3?
0.3
What is the biggest value in 274702, -0.3, -1/4?
274702
Which is the second biggest value? (a) -10 (b) -3935 (c) -5 (d) -0.4 (e) 5/4
d
Which is the biggest value? (a) 5 (b) 3 (c) 2381/6
c
What is the fourth biggest value in 5, -8/9, -0.11, -0.18, -4?
-8/9
What is the third biggest value in 47, -18/11, 2/1085?
-18/11
Which is the fifth biggest value? (a) -114/5 (b) -3 (c) 1/15 (d) 2/15 (e) 1/2
a
What is the third smallest value in 1/3, 5, 2/7, -2.2?
1/3
Which is the smallest value? (a) 97 (b) -0.12 (c) -1/6 (d) 1 (e) 0.2
c
What is the third smallest value in 1, 2/5, -2, 0.1509, 8?
2/5
Which is the third smallest value? (a) 1 (b) 0.5 (c) 608/7 (d) -8/5
a
Which is the third biggest value? (a) 0 (b) 1/3 (c) -5 (d) -18 (e) 15
a
What is the smallest value in -2/7, -2, -6, -5, -4/21?
-6
Which is the fifth biggest value? (a) -33 (b) 1/3 (c) 15 (d) -3/8 (e) 37
a
What is the smallest value in -5/2, -33/7, -5, -37?
-37
What is the third biggest value in 5, -5, -16, -3/2, 1/2?
-3/2
Which is the third biggest value? (a) -59 (b) -5 (c) 5 (d) -8 (e) 8
b
Which is the second biggest value? (a) 4 (b) -0.248 (c) -1/11
c
Which is the second biggest value? (a) 3 (b) 185103 (c) -0.3
a
Which is the third biggest value? (a) 0 (b) -0.09 (c) -23566
c
Which is the biggest value? (a) -9 (b) -697 (c) -3
c
What is the third smallest value in -5, -3749, -1/2?
-1/2
What is the biggest value in -0.1, -8994/7, 1/14?
1/14
Which is the fourth biggest value? (a) 6/13 (b) 143 (c) -0.5 (d) -3.8
d
Which is the third biggest value? (a) -2/9 (b) 1.5 (c) 4 (d) 70.1
b
What is the third smallest value in -1, -4, -0.23, -2, 1/8, -97?
-2
Which is the sixth biggest value? (a) 2/7 (b) -1/4 (c) 0.2 (d) -0.3 (e) 20010 (f) 1/8
d
What is the smallest value in 472, 4, -1/8, 0.07?
-1/8
What is the sixth biggest value in -0.6, -4, 13, 3, -2/7, -20.4?
-20.4
What is the biggest value in 0, 24/397, 0.5, -4, -2/3?
0.5
Which is the third biggest value? (a) -1/17 (b) 0.02 (c) 6
a
Which is the second smallest value? (a) 0.15 (b) 1388 (c) 5
c
Which is the fourth smallest value? (a) 0.1 (b) -4 (c) -408898 (d) 0.07
a
What is the smallest value in -5, -0.255, -14, 0.3?
-14
What is the fourth smallest value in 2, -0.197, 1/3, -0.3, 22?
2
Which is the third biggest value? (a) -24 (b) 2/7 (c) 19.5 (d) 4/3 (e) 1
e
Which is the third biggest value? (a) 2 (b) 4 (c) -3.61 (d) 0.4
d
What is the second smallest value in -4, -3/11, 0.1, -3?
-3
Which is the biggest value? (a) -1/3 (b) -1/19 (c) -9 (d) -1/7 (e) 1/3 (f) -8
e
What is the biggest value in 1/19, 2/9, -6.85?
2/9
Which is the second smallest value? (a) 2/3 (b) 4 (c) 2864
b
Which is the biggest value? (a) -0.02 (b) -5 (c) 0.0006 (d) -1
c
Which is the fourth smallest value? (a) 0.2 (b) 0.02 (c) 129 (d) 148
d
Which is the second smallest value? (a) -2.46 (b) -2 (c) 3 (d) 0.3 (e) 4 (f) 1/20
b
Which is the biggest value? (a) -775 (b) -5 (c) 15
c
What is the second biggest value in -6, 0.3, -127, -1/5, 359?
0.3
What is the fourth smallest value in -2/79, 2, 0, -3, -1.8?
0
Which is the fifth biggest value? (a) -0.07 (b) -1 (c) -2 (d) -2/385 (e) -5/3
c
Which is the second smallest value? (a) 0.4 (b) 2/13 (c) 0.02 (d) 100 (e) 1
b
What is the fifth smallest value in -5, 0.4, -3, 29, 3/4, -2/93?
3/4
What is the third smallest value in -16609, 0.3, 0.5, -0.3?
0.3
Which is the fifth biggest value? (a) -0.8 (b) 17/4 (c) 0 (d) 0.5 (e) 2/97
a
What is the third smallest value in 192/7, 3/2, -5?
192/7
What is the third biggest value in 5, 75, 9, 3?
5
What is the fourth smallest value in 0.1, -15, -1/536, 0.4, 3/2, 2/7?
2/7
What is the second smallest value in -96, 0.2, -0.2, 3, -278, 0.09?
-96
What is the second biggest value in -37.57, -11, -0.5?
-11
What is the third smallest value in -0.3, 0.01, -820, -3?
-0.3
Which is the fourth smallest value? (a) 5 (b) 6 (c) 19 (d) 6/5 (e) -2 (f) -4
a
What is the second smallest value in -141, -4, 5/4, 2/5?
-4
Which is the smallest value? (a) 40 (b) -4 (c) -46 (d) 3 (e) -3 (f) 1/4
c
Which is the third biggest value? (a) -0.1 (b) 1 (c) -2/35 (d) -3 (e) 0.217
c
What is the biggest value in 231, 0.4, 0.14, -3, 1, -0.1?
231
Which is the third biggest value? (a) -2/225063 (b) -8 (c) 2
b
What is the sixth biggest value in -1, 166/5, -7, -3, 4, -4?
-7
Which is the fifth biggest value? (a) -1 (b) -20/29 (c) 2/7 (d) 0.1 (e) -4 (f) 30/11
a
What is the second smallest value in 1/4, 0, 5, 2/5883, -0.4?
0
Which is the biggest value? (a) 3 (b) 1 (c) 43/42 (d) 67
d
What is the smallest value in -3/29, 6, -1/2, 0.25?
-1/2
Which is the second biggest value? (a) -2/1129 (b) 5 (c) 1/46
c
What is the fourth smallest value in 4, 0.5, 5, 27, 17?
17
Which is the fourth biggest value? (a) 2/3 (b) -890 (c) -1/7 (d) -2/11 (e) 3 (f) -1
d
What is the fourth biggest value in 2, 31, 5, -1344?
-1344
What is the biggest value in 2/9, 0.79, -217?
0.79
What is the third smallest value in 0.2
|
{
"pile_set_name": "DM Mathematics"
}
|
Q:
HttpURLConnection.getInputStream() hangs and never finishes
I am attempting to get a token from a JSON endpoint using the following code:
public class FetchToken extends AsyncTask<String, Void, Void> {
String data = "";
String token = "";
public TokenDelegate delegate = null;
@Override
protected Void doInBackground(String... identity) {
try {
if (identity.length == 1) {
URL url = new URL(identity[0]);
HttpURLConnection httpURLConnection = (HttpURLConnection) url.openConnection();
InputStream inputStream = httpURLConnection.getInputStream();
BufferedReader bufferedReader = new BufferedReader(new InputStreamReader(inputStream));
String line = "";
while (line != null) {
line = bufferedReader.readLine();
data = data + line;
}
JSONObject JO = new JSONObject(data);
token = JO.get("token").toString();
}
} catch (MalformedURLException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
} catch (JSONException e) {
e.printStackTrace();
}
return null;
}
@Override
protected void onPostExecute(Void aVoid) {
Log.d("token", token);
//delegate.processFinish(token);
super.onPostExecute(aVoid);
}
}
I have tried printing to the debug log after every line, and it seems that the line InputStream inputStream = httpURLConnection.getInputStream(); never finishes, nor does the rest of the code get executed. However, this is taking place inside of an AsyncTask, and the onPostExecute method of that AsyncTask is firing as if the task has been completed, which it doesn't seem as if it has been.
None of the catch blocks are being executed, it seems like, and when I inspect the Network panel in the Android Profiler then the request says it is not passing anything back. However, accessing the JSON endpoint in browser brings everything back correctly.
Hopefully someone knows what the problem may be. I've attempted to google things to no avail. Let me know if more information is needed. Thanks.
A:
The line InputStream inputStream = httpURLConnection.getInputStream(); was never completing because it was throwing a java.security.cert.CertPathValidatorException. The error was not being logged because I was not catching this type of error in my try / catch block.
A:
Have a look at below code and modify your code accordingly.
1) if method is GET, then data or headers can be null.
Example usage
url = <some url>
Map<String, String> headers = new HashMap<String, String>();
headers.put("Content-Type", "application/json");
headers.put("Accept", "application/json");
a) GET :
invokeHTTPRequest(url, "GET", headers, null);
b) POST
Student st = new Student();
Gson gson = new Gson();
String jsonReqData = gson.toJson(st);
String resp = invokeHTTPRequest3(url, "POST", headers, jsonReqData);
public static String invokeHTTPRequest(String urlString, String method, Map headers,String data) {
URL url = null;
try {
url = new URL(urlString);
HttpURLConnection conn = null;
conn = (HttpURLConnection) url.openConnection();
conn.setRequestMethod(method);
conn.setConnectTimeout(10000);
conn.setReadTimeout(10000);
if (headers != null) {
for (Map.Entry<String, String> header : headers.entrySet()) {
conn.setRequestProperty(header.getKey(), header.getValue());
}
}
if (data != null) {
byte[] postDataBytes = data.toString().getBytes();
conn.setRequestProperty("Content-Length", String.valueOf(postDataBytes.length));
conn.setDoOutput(true);
conn.getOutputStream().write(postDataBytes);
}
BufferedReader rd = null;
if(conn.getResponseCode() == 200){
rd = new BufferedReader(new InputStreamReader(conn.getInputStream()));
}
else{
rd = new BufferedReader(new InputStreamReader(conn.getErrorStream()));
}
StringBuilder result = new StringBuilder();
String line;
while ((line = rd.readLine()) != null) {
result.append(line);
}
rd.close();
return result.toString();
} catch (Exception e) {
//handle exception here
}
}
|
{
"pile_set_name": "StackExchange"
}
|
New Home Township, Bates County, Missouri
New Home Township is one of twenty-four townships in Bates County, Missouri, and is part of the Kansas City metropolitan area within the USA. As of the 2000 census, its population was 222.
History
The township takes its name from the community of New Home, Missouri.
Geography
According to the United States Census Bureau, New Home Township covers an area of 44.63 square miles (115.6 square kilometers); of this, 44.53 square miles (115.32 square kilometers, 99.76 percent) is land and 0.11 square miles (0.28 square kilometers, 0.24 percent) is water.
Unincorporated towns
Cornland at
New Home at
Nyhart at
(This list is based on USGS data and may include former settlements.)
Adjacent townships
Charlotte Township (north)
Mount Pleasant Township (northeast)
Lone Oak Township (east)
Osage Township (southeast)
Howard Township (southwest)
Walnut Township (west)
Cemeteries
The township contains Thomas Cemetery.
Rivers
Marias des Cygnes River
Lakes
Brushy Mound Lake
School districts
Butler R-V School District
Hume R-VIII
Rich Hill R-IV
Political districts
Missouri's 4th congressional district
State House District 125
State Senate District 31
References
United States Census Bureau 2008 TIGER/Line Shapefiles
United States Board on Geographic Names (GNIS)
United States National Atlas
External links
US-Counties.com
City-Data.com
Category:Townships in Bates County, Missouri
Category:Townships in Missouri
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Developmental disorders such as autism spectrum disorders (ASD) affect nearly 14% of children in the United States. Diagnostic methods for conditions such as ASD vary considerably, and even the use of “best practice” tools provides rather poor sensitivity and specificity to the conditions. Late diagnosis of developmental disabilities reduces effectiveness of treatments and often results in poor outcomes. Furthermore, treatment providers (e.g., pediatricians or other medical professionals) lack adequate tools for measuring progress in these conditions.
|
{
"pile_set_name": "USPTO Backgrounds"
}
|
Monday April 11 kicked off FLEFF Week on the campus of Ithaca College. Our first event was Checkpoints: New Media Art Installation Center, moderated by Tim Murray from Cornell University. Artists showcasing their work were Renate Ferrow of Cornell University and Megan Roberts and Raymond Ghirardo of Ithaca College.
Tim Murray gave a wonderful presentation on the Experimental Television Center (ETC) in the beginning of the program. I want to take a moment and educate everyone on this great artistic haven that will soon be no more come the end of this May.
Monday April 11 kicked off FLEFF Week on the campus of Ithaca College. Our first event was Checkpoints: New Media Art Installation Center, moderated by Tim Murray from Cornell University. Artists showcasing their work were Renate Ferrow of Cornell University and Megan Roberts and Raymond Ghirardo of Ithaca College.
Tim Murray gave a wonderful presentation on the Experimental Television Center (ETC) in the beginning of the program. I want to take a moment and educate everyone on this great artistic haven that will soon be no more come the end of this May.
Blog was written by Kelsey Greene, Documentary Studies and Production, '13, FLEFF intern, Buffalo, New York
I am not usually the kind of student who pulls out their computer during class to go online, but I have been instructed to do so for this blog.
I am sitting in the third row of seats in room 115 of the Center for Health Sciences building. Renate Ferro is currently speaking to the audience of mostly students who regularly attend Dr. Zimmerman's class Nonfiction Film Theory at this time.
After an introduction from Dr. Zimmerman, Tim Murray began the discussion with a brief history of video art and new media. I found out a great deal started right here in Ithaca, NY!
Blog was written by Kelsey Greene, Documentary Studies and Production, '13, FLEFF intern, Buffalo, New York
I am not usually the kind of student who pulls out their computer during class to go online, but I have been instructed to do so for this blog.
I am sitting in the third row of seats in room 115 of the Center for Health Sciences building. Renate Ferro is currently speaking to the audience of mostly students who regularly attend Dr. Zimmerman's class Nonfiction Film Theory at this time.
After an introduction from Dr. Zimmerman, Tim Murray began the discussion with a brief history of video art and new media. I found out a great deal started right here in Ithaca, NY!
Blog was written by Kelsey Greene, Documentary Studies and Production, '13, FLEFF intern, Buffalo, New York
Tomorrow there will be an incredible event that every one should try and attend. The event is called New Media Art Installation and will be held at 3 p.m. in room 115 in the Center for Natural Sciences. The event will be free, which makes it even better! It will be moderated by Timothy Murray and feature artists Renate Ferro, Megan Roberts, and Raymond Ghirardo. This special event is also part of the celebration of the 40th anniversary of the Experimental Television Center. Make sure you don't miss out on this spectacular opportunity!...
]]>
Blog was written by Kelsey Greene, Documentary Studies and Production, '13, FLEFF intern, Buffalo, New York
Tomorrow there will be an incredible event that every one should try and attend. The event is called New Media Art Installation and will be held at 3 p.m. in room 115 in the Center for Natural Sciences. The event will be free, which makes it even better! It will be moderated by Timothy Murray and feature artists Renate Ferro, Megan Roberts, and Raymond Ghirardo. This special event is also part of the celebration of the 40th anniversary of the Experimental Television Center. Make sure you don't miss out on this spectacular opportunity!...
|
{
"pile_set_name": "Pile-CC"
}
|
Messiah (American TV series)
Messiah is an American thriller web television series created by Michael Petroni. The first season consists of ten episodes, which were released on Netflix on January 1, 2020. The series stars Mehdi Dehbi, Tomer Sisley, Michelle Monaghan, John Ortiz, Melinda Page Hamilton, Stefania LaVie Owen, Jane Adams, Sayyid El Alami, Fares Landoulsi, and Wil Traval.
Premise
The series focuses on the modern world's reaction to a man who first appears in the Middle East, whose followers claim him to be the eschatological return of 'Isa (Jesus). His sudden appearance and apparent miracles spark a growing international following, casting doubts around who he really is, a case investigated by a CIA officer.
Cast
Main
Mehdi Dehbi as Al-Masih
Tomer Sisley as Aviram Dahan
Michelle Monaghan as CIA Case Officer Eva Geller
John Ortiz as Felix Iguero
Melinda Page Hamilton as Anna Iguero
Stefania LaVie Owen as Rebecca Iguero
Jane Adams as Miriam Keneally
Sayyid El Alami as Jibril Medina
Fares Landoulsi as Samir
Wil Traval as Will Mathers
Recurring
Philip Baker Hall as Kelman Katz
Beau Bridges as Edmund DeGuilles
Hugo Armstrong as Ruben
Barbara Eve Harris as Katherin
Nimrod Hochenberg as Israel
Emily Kinney as Staci Kirmani
Jackson Hurst as Jonah Kirmani
Nicole Rose Scimeca as Raeah Kirmani
Ori Pfeffer as Alon
Rona-Lee Shim'on as Mika Dahan
Kenneth Miller as Larry
Assad Bouab as Qamar Maloof
Dermot Mulroney as President John Young
Episodes
Production
Development
On November 17, 2017, it was announced that Netflix had given the production a series order for a first season consisting of ten episodes. The series was created by Michael Petroni who is also credited as an executive producer and lead showrunner of the series. Additional executive producers include Andrew Deane, James McTeigue, Mark Burnett and Roma Downey. Production companies involved with the series include Industry Entertainment and LightWorkers Media.
Casting
In January 2018, it was announced that John Ortiz, Tomer Sisley and Mehdi Dehbi would star in the series. In May 2018, it was announced that Michelle Monaghan had been cast in a starring role. In June 2018, it was reported that Melinda Page Hamilton, Stefania LaVie Owen, Jane Adams, Sayyid El Alami, Fares Landoulsi and Wil Traval had joined the main cast. In the same month, it was announced that Beau Bridges and Philip Baker Hall had joined the cast in a recurring capacity.
Filming
Principal photography for the first season took place in Amman, Jordan; Albuquerque, Mountainair, Estancia, Belen, Santa Fe and Clines Corners, New Mexico from June 2018 to August 2018.
Release
The first season was released on January 1, 2020.
Marketing
On December 3, 2019, the official trailer for the series was released by Netflix.
Reception
Critical reception
The review aggregator website Rotten Tomatoes reported a 43% approval rating for the first season, based on 23 reviews. The website's critical consensus states, "A promising premise and superb ensemble can't save Messiah from its own bland storytelling."
Controversy
The trailer received negative reception from some Muslim audiences. In December 2019, it was announced in a press conference that The Royal Film Commission of Jordan requested Netflix to refrain streaming Messiah in the country due to the provocative subject matter and controversial religious content covered in the series.
References
External links
(Official site)
Category:2020 American television series debuts
Category:American action television series
Category:Central Intelligence Agency in fiction
Category:English-language television programs
Category:Fictional depictions of the Antichrist
Category:Netflix original programming
Category:Serial drama television series
Category:Suspense television series
Category:Television controversies in the United States
Category:Television series by MGM Television
Category:Thriller television series
Category:Thriller web series
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Dual-stressor selection alters eco-evolutionary dynamics in experimental communities.
Recognizing when and how rapid evolution drives ecological change is fundamental for our understanding of almost all ecological and evolutionary processes such as community assembly, genetic diversification and the stability of communities and ecosystems. Generally, rapid evolutionary change is driven through selection on genetic variation and is affected by evolutionary constraints, such as tradeoffs and pleiotropic effects, all contributing to the overall rate of evolutionary change. Each of these processes can be influenced by the presence of multiple environmental stressors reducing a population's reproductive output. Potential consequences of multistressor selection for the occurrence and strength of the link from rapid evolution to ecological change are unclear. However, understanding these is necessary for predicting when rapid evolution might drive ecological change. Here we investigate how the presence of two stressors affects this link using experimental evolution with the bacterium Pseudomonas fluorescens and its predator Tetrahymena thermophila. We show that the combination of predation and sublethal antibiotic concentrations delays the evolution of anti-predator defence and antibiotic resistance compared with the presence of only one of the two stressors. Rapid defence evolution drives stabilization of the predator-prey dynamics but this link between evolution and ecology is weaker in the two-stressor environment, where defence evolution is slower, leading to less stable population dynamics. Tracking the molecular evolution of whole populations over time shows further that mutations in different genes are favoured under multistressor selection. Overall, we show that selection by multiple stressors can significantly alter eco-evolutionary dynamics and their predictability.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Image copyright Getty Images Image caption Foreign workers in South Africa have been targeted
Nigerian President Muhammadu Buhari is visiting South Africa this week to promote ties between Africa's two largest economies.
The trip comes at a time when relations have deteriorated following a rise in attacks on foreigners in South African cities, including against Nigerians.
In response, hundreds of Nigerians have left South Africa fearful of the violence and Mr Buhari is expected to address a meeting of some of those still living in the country.
So how common are attacks against migrants and are they on the rise?
Attacks against foreigners
The South African government does not collect data on attacks or threats against foreign nationals.
However, the African Centre for Migration & Society (ACMS) has monitored these attacks across South Africa since 1994. Its Xenowatch tracker collates media reports as well as information from activists, victims and observers.
Threats, attacks and killings against foreigners in South Africa. . 2019 is the latest available data.
Violent attacks peaked in 2008 and again in 2015.
Data for 2019 (to late September) already shows that the number of attacks is approaching the level of 2015.
In 2008, there was a wave of attacks across the country against refugees and migrants - more than 60 people were reported to have been killed and thousands displaced.
In 2015, there were outbreaks of violence against non-South Africans, mostly in the cities of Durban and Johannesburg, which led to the deployment of the army to deter further unrest.
In March, the government launched an initiative to raise public awareness and improve access to services for victims of discrimination.
Human rights groups welcomed it, but said that the government needed to publicly recognise attacks on foreigners as xenophobic.
In a statement published in October 2018, South Africa's main opposition party, the Democratic Alliance, blamed the governing ANC party for a "scourge of xenophobic violence".
Image copyright Getty Images Image caption Minister of Human Settlements, Water & Sanitation, Lindiwe Sisulu, has urged the police to act against people targeting foreigners
Where are the migrants from?
About 70% of foreigners in South Africa come from neighbouring Zimbabwe, Mozambique and Lesotho.
The remaining 30% is made up of people from Malawi, UK, Namibia, eSwatini, previously known as Swaziland, India and other countries.
There are an estimated 3.6 million migrants in the country, a spokesperson for South Africa's national statistics body told the BBC, out of an overall population of well over 50 million.
Image copyright Getty Images Image caption The wave of attacks has prompted a number of initiatives to tackle discrimination
How do different regions compare?
Gauteng province, which includes South Africa's largest city Johannesburg and the capital Pretoria, has the highest rate of violence against foreign nationals, followed by the Western Cape, according to the ACMS. KwaZulu-Natal, where Durban is situated, is third.
Attacks have mainly taken place in large cities, but they have also been reported in smaller towns and rural areas.
The violence is often triggered by local disputes, with migrants being accused of taking jobs away from South Africans.
Foreign-run shops have been looted and destroyed.
South Africa attacks. . .
The country has experienced poor economic performance, with officially recorded unemployment at more than 27% at the end of last year.
And more widely, the country has one of the highest murder rates in the world .
"The causes are poverty and has its roots in apartheid," says Sharon Ekambaram, who runs the refugee and migrant rights programme for Lawyers for Human Rights.
What do you want BBC Reality Check to investigate? Get in touch
Read more from Reality Check
Follow us on Twitter
|
{
"pile_set_name": "OpenWebText2"
}
|
Electrophysiological follow-up of acute and chronic experimental allergic encephalomyelitis in the Lewis rat.
Cortical somatosensory evoked potentials (c-SEP) and flash visual evoked potentials (f-VEP) were serially recorded in acute monophasic and chronic relapsing experimental allergic encephalomyelitis (EAE) in the Lewis rat. In acute EAE, a significantly delayed latency and broadened peak of the c-SEP were observed corresponding to the clinical onset, and then returned to normal with the disappearance of clinical signs. In chronic EAE, the c-SEP showed the same changes as in acute EAE, also reflecting the first attack, remission and relapsing phase. However, chronic EAE, when paralysis had recovered in the relapsing phase, showed c-SEP abnormalities suggestive of subclinical active lesions. In contrast, the f-VEP showed no obvious abnormalities in acute or chronic EAE. These findings suggest that the c-SEP is an objective and sensitive index for detecting clinical and pathological changes in acute and chronic EAE in the Lewis rat.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
**(See the Editorial Commentary by van der Werf on pages 835--6.)**
Buruli ulcer (BU), a neglected tropical disease caused by *Mycobacterium ulcerans,* is a chronic skin and soft tissue infection that can lead to permanent disfigurement and disability. It is a poorly understood mycobacterial disease, mostly affecting rural populations in West and Central Africa \[[@CIT0001], [@CIT0002]\].
In most BU-endemic settings, a diagnosis is made on clinical and epidemiological grounds, after which treatment with an 8-week course of rifampicin plus streptomycin or clarithromycin is initiated empirically \[[@CIT0003]\], pending microbiological confirmation where available. BU presents with a diverse range of clinical signs ranging from nodules or edematous plaques to (painless) ulcerations and, rarely, bone involvement. Every clinical expression of BU can be mistaken for another skin condition \[[@CIT0004]\]. Moreover, patients who are not microbiologically confirmed respond better to treatment with antimycobacterial antibiotics, which also cover other bacteria \[[@CIT0005]\].
Buruli ulcer incidence is currently declining in several endemic countries such as Benin, Ivory Coast, and Democratic Republic of Congo. In 2016, 1864 BU patients were reported to the World Health Organization (WHO) by 10 countries, whereas \>5000 cases were reported in 2009 \[[@CIT0006]\]. It has been suggested that this decline is, at least in part, due to the introduction of control strategies \[[@CIT0006]\]. The decline in incidence resulted in a decrease in BU lesions relative to non-BU lesions treated in BU facilities \[[@CIT0007]\]. Clinical expertise in the recognition of BU is thus likely to wane, potentially resulting in more diagnostic misclassification. Therefore, laboratory confirmation of BU becomes essential as this would allow improved patient management by initiating more specific therapy in non-BU patients and limiting the prolonged course of rifampicin and streptomycin/clarithromycin to those who truly have BU.
Currently, laboratory diagnostics for BU include culture, direct smear examination (DSE) for acid-fast bacilli, histopathology, and polymerase chain reaction (PCR) targeting the insertion element IS*2404* \[[@CIT0008], [@CIT0009]\]. Studies on the accuracy of diagnostic tests of BU have used varying reference standards including histopathology \[[@CIT0010]\], PCR \[[@CIT0011], [@CIT0012]\], at least 1 positive test result \[[@CIT0013]\], a composite reference standard of several tests \[[@CIT0012], [@CIT0014]\], and latent class analysis (LCA) \[[@CIT0015]\]. This diversity in reference standards makes it difficult to summarize findings on diagnostic test accuracy \[[@CIT0008]\].
The WHO recommends that clinically suspected patients are confirmed by at least 1 of the above-mentioned laboratory tests \[[@CIT0016]\]. Others have proposed a stepwise approach starting with DSE, followed by PCR only in DSE-negative patients to reduce costs \[[@CIT0017]\]. Another alternative is a Buruli score based on clinical signs and demographic patient characteristics \[[@CIT0015]\], where only patients with an intermediate score are assessed by PCR.
An accurate diagnostic algorithm that is cost-effective, reduces time to diagnosis, and is feasible in the remote and resource-limited settings where BU is endemic has great potential to reduce misclassification and improve the management of patients presenting with BU-like skin lesions. We therefore aimed to determine the diagnostic accuracy of clinical and microbiological signs in consecutively recruited patients presenting with lesions clinically compatible with BU in a BU-endemic, low-income setting.
METHODS {#s5}
=======
Study Setting and Study Population {#s6}
----------------------------------
Consecutive patients with lesions compatible with BU (ulcers, nodules, edema or plaques) presenting between March 2012 and March 2015 at the Centres de Dépistage et de Traitement de l**'**Ulcère de Buruli (CDTUB) of Allada and Lalo and in 10 health posts of the commune of Zè in southern Benin, living in BU endemic villages, were eligible for study participation. Eligible patients identified at the health posts were referred to the CDTUB for further assessment. Patients presenting with a recent (\<2 weeks) wound of obvious noninfectious origin (eg, trauma) or with a wound of \>2 weeks with normal healing were excluded from the study.
Data Collection {#s7}
---------------
All patients were assessed clinically by a triage nurse at the participating CDTUBs or the supervising nurse at the health posts. The nurses' classification as BU or non-BU based on the WHO diagnostic criteria (young age, residence in an endemic area, undermined edges, location on limbs, necrosis, absence of pain, adenopathy or fever, and hyperpigmented edges) was verified by a clinician, based on a combination of the WHO criteria and clinical experience \[[@CIT0003]\].
Samples for mycobacterial analysis were collected at the CDTUB sites. From each ulcerated lesion, 2 swabs were taken from the undermined edges. From nonulcerative lesions, 2 fine-needle aspirates were taken from the center of the lesion. One sample was processed immediately for DSE after Ziehl-Neelsen (ZN) staining. The second sample was placed in a semisolid transport medium \[[@CIT0011]\], and stored at 4°C until weekly shipment to the Mycobacteriology Reference Laboratory (LRM) in Cotonou, Benin. At the LRM, technicians performed DSE after auramine staining using fluorescence microscopy \[[@CIT0020]\], IS*2404* real-time PCR \[[@CIT0021]\], and in vitro culture for *M. ulcerans* \[[@CIT0011], [@CIT0016]\].
One 4-mm punch biopsy was taken from every lesion and stored in 10% formalin until embedding in paraffin. One section was stained by hematoxylin-eosin and 1 section by ZN. Histopathological reading was done at the University of Parakou (Benin) and the hospital of Chambéry (France). Both histopathologists were blinded to clinical information (except for age, gender, and type and localization of the lesion) and the results of other diagnostic tests. A standardized reading form was used to score histological changes ([Supplementary Data 1](#sup1){ref-type="supplementary-material"}). Based on the score, specimens were classified as probable BU (score 7**--**15), compatible with BU (score 4--6), or not compatible with BU (score ≤3). Discordant classifications between the 2 histopathologists were reread and discussed during face-to-face meetings to reach consensus scores that were used in the analysis.
Data Management {#s8}
---------------
All data were coded and registered in a Microsoft Access database. A dedicated nurse registered all clinical data while a dedicated laboratory technician registered all laboratory data.
A random selection of the DSE readings at both CDTUBs was controlled quarterly by the LRM, which participated in the external quality assurance program for *M. ulcerans* PCR organized by the Antwerp Institute of Tropical Medicine (ITM) \[[@CIT0022]\].
Data Analysis {#s9}
-------------
### Index Tests {#s10}
The accuracy of 6 index tests was evaluated: clinical diagnosis verified by the treating clinician, DSE after ZN staining, DSE after auramine staining, PCR, culture, and histopathology.
### Reference Standard {#s11}
Because a gold standard without error or uncertainty is not available for BU \[[@CIT0023], [@CIT0024]\], the accuracy of each test was estimated using an expert panel approach in the primary analysis \[[@CIT0024]\], and PCR, the best currently available diagnostic test, in a secondary analysis \[[@CIT0015]\].
The expert panel approach was based on a stepwise evaluation of all clinical information and laboratory results. The expert panel consisted of 8 study team members and 3 independent dermatologists. The expert panel determined the final diagnosis (status) of each patient in 3 steps. In the first step, patients were classified as confirmed BU (positive by PCR and/or histopathology \[score ≥ 7\]), possible BU (PCR-negative cases that were positive by DSE, clinically suspected of BU, or had a histopathology score between 4 and 6), or non-BU (negative by all microbiological tests, clinically not suspected of BU and with a histopathology score ≤ 3). In the second step, the expert panel reviewed all patient files classified as possible or non-BU in step 1. Each expert panel member independently made a differential diagnosis for every patient based on all available clinical, demographic, epidemiological, histological and mycobacteriological information ([Supplementary Data 2](#sup1){ref-type="supplementary-material"}) as well as clinical photographs (missing for 31 patients). In the third step, all cases for whom there was disagreement regarding the classification in step 2 were discussed during 1 of 3 face-to-face meetings until consensus was reached. During these discussions, more weight was given to the opinion of the independent dermatologists compared to those panel members who were also part of the study team. Participants for whom the expert panel failed to reach a consensus were classified as BU.
### Accuracy of Diagnostic Indicators {#s12}
The accuracy estimates were calculated as sensitivities, specificities, positive predictive value (PPV), and negative predictive value (NPV). Combined accuracy estimates were also calculated. For histopathology, the accuracy estimates were determined at the cutoff scores for compatible with BU (score 4) and probable BU (score 7) and at the score with an optimal combination of sensitivity and specificity on the receiver operating characteristic (ROC) curve.
The accuracy of 3 diagnostic approaches was then estimated: the WHO recommendation to consider patients who are positive by at least 1 laboratory test as confirmed BU; the stepwise approach to reserve PCR to DSE-negative patients; and the clinical Buruli score that is followed by PCR only when patients have an intermediate score \[[@CIT0015]\] ([Supplementary Data 3](#sup1){ref-type="supplementary-material"}). Three of the characteristics that make up the Buruli score (yellow and green color, and lesion hyposensitivity) were not available in our dataset. The Buruli score was only tested on ulcerative lesions as 2 of its components (characteristic smell and undermined edges) are unavailable for nonulcerative lesions.
Next, since diagnostic accuracy estimates can vary across patient subgroups, we evaluated effect modification by study site, type of lesion, recruitment type, time since start of study, availability of photographs, patient delay before consultation, transport time of samples to the LRM, and HIV status.
Statistical Analysis {#s13}
--------------------
All analyses were performed in R version 3.3.2. through RStudio version 1.0.136.
The standards for reporting diagnostic studies (STARD \[[@CIT0027]\]) were followed while writing this manuscript ([Supplementary Data 14](#sup1){ref-type="supplementary-material"}).
Ethical Considerations {#s14}
----------------------
The study protocol was approved by the Provisional National Committee for Ethics in Health Research in Benin, the Institutional Review Board of ITM, and the Ethical Committee of the University Hospital of Antwerp. The study also received administrative authorization of the Benin Ministry of Health Ethics Board. All participants gave written informed consent.
RESULTS {#s15}
=======
Participants {#s16}
------------
Between March 2012 and March 2015, 260 eligible patients presented at the study sites: 166 with symptoms or signs clinically compatible with BU and 94 with lesions clinically not compatible with BU ([Figure 1](#F1){ref-type="fig"}).
{#F1}
Of the 166 patients clinically suspected to have BU, 133 were enrolled. The 33 not recruited did not differ demographically nor clinically from the study participants, suggesting that, at least among those clinically suspected to have BU, there was no overt selection bias ([Supplementary Data 4](#sup1){ref-type="supplementary-material"}). The total number of eligible patients with lesions clinically not compatible with BU was not documented.
Among the 227 participants recruited, 205 had complete test results and were included in the analysis ([Figure 1](#F1){ref-type="fig"}). Their demographic and clinical characteristics are summarized in [Table 1](#T1){ref-type="table"}. The patients included in the analysis did not differ demographically nor clinically from the ones with incomplete data ([Supplementary Data 5](#sup1){ref-type="supplementary-material"}).
######
Demographic and Clinical Characteristics of the 205 Study Participants
Characteristic All Patients BU Patients (n = 131) Non-BU Patients (n = 74) OR (95% CI) *P* Value
----------------------------- -------------- ----------------------- -------------------------- --------- ---------- ----------- -------- ----------------- ----------
Female sex 89 (43.41) 62 (47.33) 27 (36.49) 1.56 (.87--2.81) .13
Age, y, median (IQR) 19 (9--42) 12 (9--42) 40 (9--42) \<.00001
CDTUB 113 (55.12) 51 (38.93) 62 (83.78) 0.09 (.04--.18) \<.00001
Allada 100 50 50
Lalo 13 1 12
HIV status 0.74 (.17--3.23) .69
Infected 8 (5.52) 5 (5.00) 3 (6.67)
Uninfected 137 (94.48) 95 (95.00) 42 (93.33)
Not tested 60 31 29
Clinically BU, WHO category 127 (61.95) 120 (91.60) 7 (9.46) 104.42 (38.66--282.04) \<.00001
1 25 (19.69) 22 (18.33) 3 (42.86)
2 69 (54.33) 65 (54.17) 4 (57.14)
3 33 (25.98) 33 (27.50) 0
Clinically non-BU 78 (38.05) 11 (8.40) 67 (90.54)
Necrotizing fasciitis 36 (46.15) 5 (45.45) 31 (46.27)
Chronic ulcer 22 (28.21) 2 (18.18) 20 (29.85)
Abscess 5 (6.41) 5 (7.46)
Infected wound 3 (3.85) 1 (9.09) 2 (2.99)
Chronic traumatic wound 2 (2.56) 2 (2.99)
Tumor 2 (2.56) 1 (9.09) 1 (1.49)
Erysipelas 1 (1.28) 1 (1.49)
Osteomyelitis 1 (1.28) 1 (9.09)
Kaposi sarcoma 1 (1.28) 1 (1.49)
Varicose ulcer 1 (1.28) 1 (1.49)
Necrotic wound 1 (1.28) 1 (9.09)
Suppuration 1 (1.28) 1 (1.49)
Cervical adenopathy 1 (1.28) 1 (1.49)
Ganglionary ulcer 1 (1.28) 1 (1.49)
Data are presented as No. (%) unless otherwise indicated.
Abbreviations: BU, Buruli ulcer; CDTUB, Centres de Dépistage et de Traitement de l**'**Ulcère de Buruli; CI, confidence interval; HIV, human immunodeficiency virus; IQR, interquartile range; OR, odds ratio; WHO, World Health Organization.
Expert Panel Approach {#s17}
---------------------
The expert panel approach classified 131 (64%) patients as BU and 74 (36%) as non-BU ([Figure 1](#F1){ref-type="fig"}). Among the 74 non-BU patients, the expert panel classified 30 as non-BU with confirmed differential diagnoses and 44 as non-BU with unclear differential diagnoses ([Supplementary Data 6](#sup1){ref-type="supplementary-material"}). Among the 11 patients for whom the expert panel failed to reach a consensus and who were classified as BU, 10 were treated with an 8-week course of rifampicin and streptomycin. One patient was not treated as BU and received ciprofloxacin and cloxacillin.
Accuracy of Diagnostic Indicators {#s18}
---------------------------------
The results of the index tests are summarized in [Supplementary Materials 7](#sup1){ref-type="supplementary-material"} and [8](#sup1){ref-type="supplementary-material"}.
Using the expert panel as the reference, clinical diagnosis by the treating clinician had the highest sensitivity (0.92 \[95% confidence interval {CI}, .85--.96\]), followed by PCR (0.65 \[95% CI, .56--.73\]), Ziehl-Neelsen DSE (0.47 \[95% CI, .38--.55\]), auramine DSE (0.28 \[95% CI, .21--.37\]), and culture (0.28 \[95% CI, .20--.36\]) ([Figure 2](#F2){ref-type="fig"}). The specificities and PPV of all diagnostics were high (≥0.91 and ≥0.92, respectively). The NPV of the clinical diagnosis was higher (0.86 \[95% CI, .76--.93\]) than that of any of the other tests (≤0.62 for all other tests). Using PCR as the reference in a secondary analysis, clinical diagnosis remained the diagnostic method with the highest sensitivity (0.92 \[95% CI, .84--.97\]) ([Figure 2](#F2){ref-type="fig"}).
{#F2}
For histopathology, the sensitivity and specificity at a cutoff score of 4 (compatible with BU) were 0.54 (95% CI, .45--.63) and 0.73 (95% CI, .61--.83), respectively. At cutoff, 7 (probable BU), the sensitivity and specificity were 0.18 (95% CI, .12--.26) and 1.00 (95% CI, .95--1.00). Using ROC curve analysis, we could not identify a better cutoff given the low combination of sensitivity and specificity across the entire range of scores, with a maximum performance (optimal combination of sensitivity and specificity) at a score of 3.5 (sensitivity: 0.80 \[95% CI, .73--.86\]; specificity: 0.57 \[95% CI, .45--.68\]). The area under the curve was 0.73 (95% CI, .67--.80; [Supplementary Data 9](#sup1){ref-type="supplementary-material"}).
The accuracy of different combinations of diagnostic tests resulted in an increased sensitivity only when PCR was included without reducing specificity nor PPV ([Figure 3](#F3){ref-type="fig"}).
{#F3}
When evaluating effect modification by important covariates, only the performance of clinical diagnosis differed according to study site and type of lesion, with a lower specificity in decentralized settings, and a higher specificity in ulcerative lesions ([Supplementary Data 10](#sup1){ref-type="supplementary-material"}). The NPV of both DSE assays, PCR, and culture was lower among patients recruited in decentralized sites, reflecting the higher BU prevalence in decentralized sites.
In a sensitivity analysis, no significant differences were observed in accuracy estimates when classifying the 11 participants for whom the expert panel could not reach a consensus as non-BU ([Supplementary Data 11](#sup1){ref-type="supplementary-material"} and [12](#sup1){ref-type="supplementary-material"}).
Assuming that sensitivity and specificity of clinical diagnosis and microbiological tests remain constant, a further reduction of the proportion of BU patients would impact the predictive values, with increases in NPV and decreases in PPV ([Supplementary Data 13](#sup1){ref-type="supplementary-material"}).
Accuracy of Published Diagnostic Algorithms {#s19}
-------------------------------------------
All evaluated diagnostic algorithms had similar performances ([Figure 4](#F4){ref-type="fig"}).
{#F4}
DISCUSSION {#s20}
==========
In a BU-endemic setting, trained clinicians clinically diagnose BU with a sensitivity of 92% (95% CI, 85%--96%) and specificity of 91% (95% CI, 81%--96%). The sensitivity of a clinical diagnosis was higher than the sensitivity of any laboratory test. Typical ulcerated BU lesions have indeed been reported to be easily diagnosed clinically in BU-endemic areas by experienced clinicians \[[@CIT0016]\]. Despite the excellent performance of clinical diagnosis, 14% of the study participants clinically not suspected to have BU were reclassified as BU patients by the expert panel procedure, suggesting that there may be a nonnegligible level of underdiagnosis. The majority (64%) of these missed diagnoses were positive by PCR. In clinical practice, these patients do not receive any laboratory test and are not treated for BU unless their clinical evolution suggests BU, resulting in a considerable delay to treatment. However, if the burden of BU would continue reducing, it would virtually never be missed by the clinicians, but a considerable proportion of clinical suspects would not have BU ([Supplementary Data 13](#sup1){ref-type="supplementary-material"}).
The sensitivity of PCR (65% \[95% CI, 56%--73%\]) in the present study was lower than generally reported in studies using a variety of reference standards (85.4% \[[@CIT0013]\], 86.0% \[[@CIT0028]\], 87.5% \[[@CIT0014]\], 100% \[[@CIT0015]\]). Clinical characteristics were included into the expert panel approach which resulted in a considerable proportion of PCR-negative patients being classified as BU and therefore more false-negative PCR results than in studies using only microbiological assays as reference standards. However, the addition of PCR to DSE resulted in an increased sensitivity (69% \[95% CI, 60%--77%\]) while the combined specificity remained high (93% \[95% CI, 85%--98%\]).
The performance of published diagnostic algorithms for BU in endemic settings was similar. However, the Buruli score---based on clinical and demographic patient characteristics, reserving PCR for patients with an intermediate score \[[@CIT0015]\]---can only be used for the diagnosis of ulcerative lesions. The stepwise approach reserving PCR to DSE-negative patients would therefore be the most cost-effective diagnostic algorithm with the shortest time to results in both ulcerative and nonulcerative lesions.
The performance of histopathology was poor in this study. Its accuracy estimates were low at both cutoff scores of 4 and 7. The ROC analysis of the histopathology consensus scores did not allow the identification of a more optimal cutoff, possibly due to a poor performance of the standardized reading form. Moreover, the pathologists received biopsies in batches blinded to other test results and clinical information. While blinding of histopathologists allows assessing the value of the test per se, it is in contrast with clinical practice where histopathology reading is interpreted taking into account clinical information. Blinding may thus have underestimated the true added value of histopathology, resulting in test review bias \[[@CIT0029]\]. An additional limitation of histopathology in the present analysis is the use of consensus scores. Moreover, the 4-mm punch biopsies recommended by the WHO may sometimes be too small to contain the typical histological changes. Also, the localization of the biopsy in the lesion determines the histological changes captured by the sample. Since in clinical practice histopathology is used when all microbiological tests are negative but there is still a clinical suspicion of BU, a cutoff should be selected at a high sensitivity. However, in the present study the sensitivity of histopathology was disappointingly low at every cutoff.
Because a gold standard without error or uncertainty is not available for BU, we used an expert panel approach. However, the expert panel had access to all available diagnostic indicators, resulting in incorporation bias \[[@CIT0023]\] and possibly overestimating the specificity and PPV of the diagnostic indicators while underestimating their sensitivity and NPV. The true accuracy estimates are probably situated between those measured by PCR and the expert panel. An alternative for diagnostic studies where a good reference standard is lacking is LCA \[[@CIT0030]\], which relates observed patterns of test results to unknown or latent categories of patients---those with and those without the condition. For BU, LCA has only been used in 1 study \[[@CIT0015]\].
Inclusion of multiple sites, clinicians, and histopathologists may have increased interobserver variability. The effect of interobserver variability is likely to be larger for the more subjective tests such as clinical diagnosis and histopathology. We deliberately choose a pragmatic approach, as recommended by Grobbee and Hoes \[[@CIT0031]\], where all diagnostic determinants are assessed as much as possible according to daily practice and by the practicing physician, with some effort to standardize measurements. However, generalizing the study results should be done with care. Health workers in other BU-endemic settings may be less experienced than those working in southern Benin.
CONCLUSIONS {#s21}
===========
Despite the excellent performance of clinical diagnosis, there may be an important level of underdiagnosis. A broader clinical suspicion is therefore recommended to reduce missed BU diagnoses allowing improved patient management.
Taking into consideration diagnostic accuracy, time to results, cost-effectiveness, and clinical generalizability, the stepwise diagnostic approach reserving PCR to microscopy-negative patients performed best and would be suitable in the remote and resource-limited settings where BU is endemic.
Supplementary Data {#s22}
==================
Supplementary materials are available at *Clinical Infectious Diseases* online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
######
Click here for additional data file.
***Acknowledgments.*** We thank Jim Zeegelaar and Felix Atadokpede for their contributions to the clinical expert panel, and Achilleas Tsoumanis, Joris Menten, and Meryam Krit for help with coding in R and advice on statistical analyses.
***Disclaimer.*** The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
***Financial support.*** This work was supported by the Medicor Foundation ([www.medicor.li](http://www.medicor.li)) and UBS Optimus Foundation ([www.ubs.com/microsites/optimus-foundation](http://www.ubs.com/microsites/optimus-foundation)) through the BU differential diagnosis project in Benin. Fondation Follereau Luxembourg ([www.ffl.lu/](http://www.ffl.lu/)) gives core support to the CDTUB in Allada. M. E. was supported by the Department of Economy, Science and Innovation ([www.ewi-vlaanderen.be/](http://www.ewi-vlaanderen.be/)) of the Flemish government (Belgium). B. d. J. was supported by the European Research Council-INTERRUPTB (starting grant number 311725).
***Potential conflicts of interest.*** All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
|
{
"pile_set_name": "PubMed Central"
}
|
Two famous quotes by President
Harry S. Truman
are linked to the date April 11.
The first is something Truman said about a historic announcement he made on April 11, 1951 .
On that date, Truman announced his decision to fire General Douglas MacArthur , the “Supreme Commander for the Allied Powers” in Korea, for disagreeing with his policy of limiting the expansion of the Korean War.
It was the culmination of a dispute that started the previous year.
In November and December of 1950, Mao Tse-tung (a.k.a. Mao Zedong), the new leader of the Communist People’s Republic of China, sent hundreds of thousands of Red Chinese troops to fight alongside North Koreans against American and South Korean forces.
MacArthur wanted to respond by attacking China, possibly with nuclear weapons.
Truman firmly squelched that idea.
But MacArthur, who’d enjoyed great popularity with the public since World War II and had a huge ego, decided to try to play a game of political chicken with Truman.
In late March, he wrote a letter to Republican Congressman Joseph W. Martin in which he clearly criticized the President’s policy and slyly played the Red scare card.
His letter suggested that China’s intervention in Korea should be met with “maximum counterforce” and said, in an obvious reference to Truman: “It seems strangely difficult for some to realize that here in Asia is where the Communist conspirators have elected to make their play for global conquest.”
As MacArthur expected, Martin made the letter public.
Truman was furious. And, after a series of discussions with the Joint Chiefs of Staff, he decided to relieve MacArthur of his command.
On the night of April 11, 1951, Truman officially announced his firing of MacArthur in a special broadcast to the American people.
Truman’s famous quote about that decision came to light years later, in 1974, with the publication of the best-selling book Plain Speaking: An Oral Biography of Harry S. Truman .
The book, by Merle Miller, was based on taped-recorded interviews made with Truman in the 1960s.
In one chapter, Miller provided Truman’s response when asked why he decided to fire General MacArthur.
Truman’s salty answer soon became a famous quote .
“I fired him because he wouldn’t respect the authority of the President…I didn’t fire him because he was a dumb son of a bitch, although he was, but that’s not against the laws for generals. If it was, half to three-quarters of them would be in jail.”
Of course, there’s another, even better-known quote that’s associated with Truman’s firing of MacArthur.
On April 19, 1951 , eight days after he was relieved of duty, MacArthur made a high-profile “farewell address” to a joint meeting of Congress. That speech included the familiar line: “Old soldiers never die, they just fade away.”
There’s also a second Truman quotation that’s linked to the date April 11th, but it’s not about MacArthur. It’s Truman’s humorous, oft-cited explanation of the difference between a politician and a statesman.
On April 11, 1958 , speaking to the Reciprocity Club in Washington, D.C., the retired president said :
“A statesman is a politician who’s been dead ten or fifteen years.”
To read the background on that famous quotation, see this previous post on This Day in Quotes .
* * * * * * * * * *
Comments? Questions? Corrections? Post them on the Famous Quotations Facebook page .
|
{
"pile_set_name": "OpenWebText2"
}
|
---
abstract: 'We investigate the FIR properties of a sample of BCDs observed by *AKARI*. By utilizing the data at wavelengths of $\lambda =65~\mu$m, 90 $\mu$m, and 140 $\mu$m, we find that the FIR colours of the BCDs are located at the natural high-temperature extension of those of the Milky Way and the Magellanic Clouds. This implies that the optical properties of dust in BCDs are similar to those in the Milky Way. Indeed, we explain the FIR colours by assuming the same grain optical properties, which may be appropriate for amorphous dust grains, and the same size distribution as those adopted for the Milky Way dust. Since both interstellar radiation field and dust optical depth affect the dust temperature, it is difficult to distinguish which of these two physical properties is responsible for the change of FIR colours. Then, in order to examine if the dust optical depth plays an important role in determining the dust temperature, we investigate the correlation between FIR colour (dust temperature) and dust-to-gas ratio. We find that the dust temperature tends to be high as the dust-to-gas ratio decreases but that this trend cannot be explained by the effect of dust optical depth. Rather, it indicates a correlation between dust-to-gas ratio and interstellar radiation field. Although the metallicity may also play a role in this correlation, we suggest that the dust optical depth could regulate the star formation activities, which govern the interstellar radiation field. We also mention the importance of submillimetre data in tracing the emission from highly shielded low-temperature dust.'
author:
- |
Hiroyuki Hirashita$^{1}$[^1] and Tomohiro T. Ichikawa$^{2}$\
$^1$ Institute of Astronomy and Astrophysics, Academia Sinica, P.O. Box 23-141, Taipei 10617, Taiwan\
$^2$ Centre for Computational Sciences, University of Tsukuba, Tsukuba 305-8577, Japan
date: 2009 March 3
title: 'Effects of dust abundance on the far-infrared colours of blue compact dwarf galaxies'
---
\[firstpage\]
dust, extinction — galaxies: dwarf — galaxies: evolution — galaxies: ISM — infrared: galaxies
Introduction
============
The far-infrared (FIR) emission from galaxies is often used to trace the star formation activities [@kennicutt98; @inoue00; @iglesias04]. Dust grains are the source of FIR emission, and the strong connection between FIR luminosity and star formation rate can be explained if the ultraviolet (UV) light from massive stars is the dominant source of dust heating. The FIR spectral energy distribution (SED) of dust grains reflects various information on the grains themselves and on the sources of grain heating [e.g. @takagi03; @takeuchi05; @dopita05]. Dust temperature is determined by the intensity of interstellar radiation field (ISRF) and the optical properties of dust (i.e. how it absorbs and emits light). Since dust grains absorb UV light efficiently, the FIR luminosity and the dust temperature mostly reflect the UV radiation field [@buat96].
Observationally, dust temperature can be estimated from FIR colour, which is defined as the flux ratio between two FIR wavelengths. If more than two ($\geq 3$) bands are available in FIR, we can take two independent FIR colours and examine a FIR colour–colour relation. A colour–colour relation, because of increased information compared with a single FIR colour, enables us to obtain not only the dust temperature but also the wavelength dependence of dust emissivity. @nagata02, adopting 60 $\mu$m, 100 $\mu$m, and 140 $\mu$m as three wavelengths, show that there is a tight relation in the FIR colour–colour diagram. This tightness implies a common wavelength dependence of FIR emissivity among various galaxies. Their work has been developed by @hibi06, who show that a tight FIR colour–colour relation found for the Milky Way dust emission, called “main correlation”, is also consistent with the FIR colour–colour relation of a sample of nearby galaxies (see also @sakon04 [@onaka07]). As stated by @hibi06, this implies that the optical properties of grains in FIR are common among the Milky Way and nearby galaxies.
Since the FIR colour–colour relation can be used to constrain the optical properties of dust grains, the FIR colour–colour relation of galaxies with various evolutionary stages provides pieces of information on the grain evolution in galactic environments. In particular, the dominant production source of dust grain could be different in early phase of galaxy evolution [@todini01; @nozawa03; @maiolino04; @schneider04]. Moreover, the effects of interstellar processing of dust grains, especially accretion of heavy elements onto dust grains, coagulation, and shattering, depend on metallicity (or dust abundance). Thus, it is probable that the grain properties evolve as galaxies evolve.
Although it is difficult to obtain the rest-frame FIR data of distant galaxies in an early phase of galaxy evolution, there are nearby possible “templates” of primeval galaxies, blue compact dwarf galaxies (BCDs). Indeed, BCDs have on-going star formation in metal-poor and gas-rich environments [@sargent70; @vanzee98], which could be similar to those in high-redshift star-forming galaxies. Moreover, BCDs harbour an appreciable amount of dust [e.g. @thuan99] and can be used to investigate the dust properties in chemically unevolved galaxies [@takeuchi05].
It is possible to investigate the metallicity dependence of dust properties by sampling BCDs with a variety of metallicity. Recently @engelbracht08 have examined the metallicity dependence of dust emission in various wavelengths by using *Spitzer* data. In FIR, they have used Multiband Imaging Photometer for *Spitzer* (MIPS) 70 $\mu$m and 160 $\mu$m bands, and have shown that there is a correlation between dust temperature derived from these two bands and metallicity. This indicates the importance of studies on dust emission in a wide metallicity range.
The *AKARI* satellite [@murakami07] provides us with a good opportunity to study FIR colour–colour relations, since Far-Infrared Surveyor (FIS) on *AKARI* has four bands in FIR (65 $\mu$m, 90 $\mu$m, 140 $\mu$m, and 160 $\mu$m) [@murakami07; @kawada07]. Indeed, seven of the eight BCDs in @hirashita08 [hereafter H08] are detected at 65 $\mu$m, 90 $\mu$m, and 140 $\mu$m. This indicates that it is possible to study the FIR colour–colour relation of BCDs by using *AKARI* data. In this paper, we add four more BCDs available in the *AKARI* archive and examine the FIR colour–colour relation of BCDs. Then, we extract information on what determines or regulates the FIR emission in metal-poor environments with the aid of theoretical models for dust emission.
This paper is organized as follows. First, in Section \[sec:model\], we describe the models of dust emission with a simple radiative transfer recipe. Then, in Section \[sec:data\], we explain the data analysis of the BCD sample observed by *AKARI*. In Section \[sec:result\], we overview the results of the model calculations in comparison with the observational data. In Section \[sec:discussion\], we discuss our results, focusing on the relation between FIR colour and dust content. Finally, the conclusion is presented in Section \[sec:conclusion\].
FIR SED Model {#sec:model}
=============
We adopt the theoretical framework of @draine01 to calculate the SED of dust emission in FIR. The framework has already been described in @hirashita07 [hereafter HHS07], but we modify it to treat BCDs in this paper. Moreover, [as described in the last paragraph in Sect. \[subsec:isrf\],]{} we newly include the effect of radiative transfer in a similar way to @galliano03. We overview our framework, focusing on the change from HHS07.
Interstellar radiation field {#subsec:isrf}
----------------------------
The stellar SEDs of BCDs are generally harder than those of spiral galaxies [@boselli03]. @madden06 also show a hard stellar radiation field from the observation of mid-infrared emission lines in low-metallicity star-forming galaxies. Thus, we change the SED of ISRF, and adopt the following fitting formula according to the averaged SED of the BCD sample in @boselli03: $$\begin{aligned}
4\pi\lambda J_\lambda^{(0)} =\left\{
\begin{array}{l}
0 ~~~~~ (\lambda_{\mu\mathrm{m}}\leq 0.0912), \\
3.02\times 10^{-4}\lambda_{\mu\mathrm{m}}^{-1.20}\\
~~~~~~~(0.0912<\lambda_{\mu\mathrm{m}}\leq 0.25), \\
1.55\times 10^{-3}\times 10^{-3}
\lambda_{\mu\mathrm{m}}^{-0.0371}\\ ~~~~~~~
(0.25<\lambda_{\mu\mathrm{m}}\leq 0.37), \\
2.39\lambda_{\mu\mathrm{m}}^{7.41}~~~~~
%~~~~~~~
(0.37<\lambda_{\mu\mathrm{m}}\leq 0.4), \\
4\pi\lambda [W_1B_\lambda (T_1)+W_2B_\lambda (T_2)]
\\ ~~~~~~~
(0.4<\lambda_{\mu\mathrm{m}}),
\end{array}
\right.\label{eq:isrf}\end{aligned}$$ where $J_\lambda^{(0)}$ is the ISRF intensity whose normalization is determined so that the intensity at wavelength $\lambda =0.2~\mu$m (almost the centre of the UV range; @buat96) is equal to the solar neighborhood ISRF estimated by @mathis83 ($4\pi\lambda J_\lambda^{(0)}$ is expressed in units of erg cm$^{-2}$ s$^{-1}$), $\lambda_{\mu\mathrm{m}}$ is the wavelength in units of $\mu$m, $B_\lambda (T)$ is the Planck function, $(T_1,\, T_2)=(1000,\, 4000~\mathrm{K})$, and $(W_1,\, W_2)=(1.61\times 10^{-15},\, 2.90\times 10^{-14})$.
In this paper, we newly include the effect of dust extinction on the ISRF. We simply assume that the ISRF intensity, $J_\lambda$, at optical depth $\tau_\lambda$ is $$\begin{aligned}
J_\lambda=J_\lambda^{(0)}\chi\exp(-\tau_\lambda)\, ,\end{aligned}$$ where $\chi$ is the scaling factor of the ISRF relative to $J_\lambda^{(0)}$ and is assumed to be independent of wavelength. For the wavelength dependence of $\tau_\lambda$, we assume the Milky Way extinction curve taken from @cardelli89 with $R_V=3.1$. Thus, the extinction at any wavelength $\lambda$, $A_\lambda$ (in units of magnitude), can be determined if we set a value of $A_V$ (extinction at $V$ band). The optical depth can be related with the extinction as $\tau_\lambda =A_V/1.086$. The following results does not change drastically if we assume an extinction curve appropriate for the Large Magellanic Clouds (LMC) or the Small Magellanic Cloud (SMC), since the detailed shape of UV–optical SED is not important in this paper. Strictly speaking, we treat scattering as effective absorption, and this treatment would overestimate the absorbed energy. Thus, we interpret $A_V$ given here as effective absorption optical depth, so that the total energy absorbed by dust in our formulation is equal to the energy that would be absorbed if scattering were properly treated.
@galliano03 also adopted a similar treatment, which is valid if dust is distributed in a thin shell surrounding the stars. Such a “screen” geometry enhances the effect of dust extinction (i.e., radiative transfer) in comparison with a mixed geometry between stars and dust. Thus, our models are suitable to examine the extent to which the radiative transfer effects could affect the FIR SEDs.
Grain properties {#subsec:properties}
----------------
We consider silicate and graphite grains in this paper. Since we are interested in the wavelength range appropriate for the *AKARI* FIS bands ($\lambda\sim 50$–180 $\mu$m; Section \[subsec:analysis\]), we neglect polycyclic aromatic hydrocarbons (PAHs), which contribute significantly to the emission at $\lambda\la 20~\mu$m [@desert90; @dwek97; @draine01].
We assume a grain to be spherical with a radius of $a$. The absorption cross section of the grain is expressed as $\pi a^2Q_\mathrm{abs}(\lambda )$, where $Q_\mathrm{abs}(\lambda )$ is called absorption efficiency. We adopt the optical constants of astronomical silicate and graphite for $\lambda <100~\mu$m from @draine84 and @weingartner01 and calculate the absorption efficiency by using Mie theory [@bohren83]. For the absorption efficiency at $\lambda >100~\mu$m, we adopt a functional form proposed by @reach95: $$\begin{aligned}
Q_\mathrm{abs}(\lambda )=
\frac{(\lambda /\lambda_0)^{-2}}
{[1+(\lambda_1/\lambda)^6]^{1/6}}\, ,\label{eq:reach}\end{aligned}$$ which behaves like $\beta =1$ ($Q_\mathrm{abs}(\lambda )\propto\lambda^{-\beta}$, where $\beta$ is called emissivity index) for $\lambda\ll\lambda_1$ and $\beta =2$ at $\lambda\gg\lambda_1$ [see also @bianchi99]. We assume that $\lambda_1=200~\mu$m, following @reach95. We set the value of $\lambda_0$ so that the continuity of $Q_\mathrm{abs}$ at $100~\mu$m is satisfied. As shown in HHS07, the wavelength dependence of $Q_\mathrm{abs}$ assumed here explains the FIR colour–colour relations of the Milky Way, the LMC and the SMC. Although the absorption efficiency at $\lambda >100~\mu$m is modified, we call the grain species “silicate” and “graphite” according to the adopted absorption efficiency at $\lambda <100~\mu$m.
The number density of grains with sizes between $a$ and $a+\mathrm{d}a$ is denoted as $n_i(a)\,\mathrm{d}a$, where the subscript $i$ denotes a grain species (silicate or graphite). Since there is little knowledge on the grain size distribution in BCDs, we simply assume a power-law form for $n_i(a)$: $$\begin{aligned}
n_i(a)=\mathcal{C}_ia^{-K}~~~(a_\mathrm{min}\leq a\leq
a_\mathrm{max})\, ,
\label{eq:size_dist}\end{aligned}$$ where $a_\mathrm{min}$ and $a_\mathrm{max}$ are the upper and lower cutoffs of grain size, respectively, and $\mathcal{C}_i$ is the normalizing constant. We assume $K=3.5$, $a_\mathrm{min}=3.5$ Å and $a_\mathrm{max}=0.25~\mu$m for both graphite and silicate [@mathis77; @li01]. As long as we treat the FIR colours of a single species (Section \[subsec:colour\]), the constant $\mathcal{C}_i$ cancels out. Because the dust composition in BCDs is still unclear, we consider silicate and graphite separately as two possible dust species.
Following HHS07, we adopt the multi-dimensional Debye models for the heat capacities of grains [@draine01]. The physical parameters necessary to characterize the heat capacities can be found in HHS07.
Calculation of FIR colours {#subsec:colour}
--------------------------
The FIR intensity $\mathcal{J}_\nu^i(\lambda )$ (per frequency per hydrogen nucleus per solid angle) of the emission from dust species $i$ at a wavelength $\lambda$ can be estimated as $$\begin{aligned}
\mathcal{J}_\nu^i(\lambda )=
\int_{a_\mathrm{min}}^{a_\mathrm{max}}\mathrm{d}
a\frac{1}{n_\mathrm{H}}n_i(a)\pi a^2Q_\mathrm{abs}
(\lambda )\int_0^\infty\mathrm{d}T\, B_\nu (T)
\frac{\mathrm{d}P_i}{\mathrm{d}T},\hspace{-1cm}\nonumber\\\end{aligned}$$ where $B_\nu (T)$ is the Planck function at frequency $\nu$ and temperature $T$, and $\mathrm{d}P_i/\mathrm{d}T$ is the temperature distribution function of the grains under the ISRF evaluated in Section \[subsec:isrf\]. The calculation method of $\mathrm{d}P_i/\mathrm{d}T$ is based on the concept of stochastic heating, and we follow the formalism of @draine01 (see their section 4). We assume that the FIR radiation is optically thin. Thus, after the radiative transfer effect, the intensity at $\lambda$ per frequency per hydrogen nucleus per solid angle, $I_\nu (\lambda )$, becomes $$\begin{aligned}
I_\nu (\lambda)=\int_0^{s}\mathcal{J}_\nu\,\mathrm{d}s\, ,\end{aligned}$$ where $s$ is the path length, which is proportional to $A_V$. Here the colour at wavelengths of $\lambda_1$ and $\lambda_2$ is defined as $I_\nu (\lambda_1)/I_\nu (\lambda_2)$, which is called $\lambda_1-\lambda_2$ colour and denoted as $(\lambda_1/\lambda_2)_\mathrm{cl}$ (we omit $\mu$m in this expression; e.g. $(60/100)_\mathrm{cl}$ and $(140/100)_\mathrm{cl}$). Since we only consider colours (i.e. flux ratios), the proportionality constant between $s$ and $A_V$ does not matter in this paper. Thus, practically, we can replace $s$ with $A_V$.
[ Since we calculate colours in this paper, we can define the colour with $A_V=0$ as the optically thin extreme ($\lim_{A_V\to 0}(\lambda_1/\lambda_2)_\mathrm{cl}$). In other words, the FIR colour calculated with $A_V=0$ (in fact $A_V\to 0$) represents the case where the dust is illuminated by a radiation field without extinction. In the following, $A_V=0$ means that the dust is illuminated by a radiation field without extinction, not that there is no dust. ]{}
Data {#sec:data}
====
Analysis of the *AKARI* data {#subsec:analysis}
----------------------------
H08 analysed eight BCDs observed by FIS onboard *AKARI* with four photometric bands of *N60*, *WIDE-S*, *WIDE-L*, and *N160*, whose central wavelengths are 65, 90, 140, and 160 $\mu$m with effective band widths of $\Delta\lambda =21.7$, 37.9, 52.4, and 35.1 $\mu$m, respectively [@kawada07]. The measured FWHMs of the point spread function are $37\pm 1''$, $39\pm 1''$, $58\pm 3''$, and $61\pm 4''$, respectively, for the above four bands. We exclude Mrk 36 from the sample, since it is not clearly detected at *WIDE-L* (140 $\mu$m). The other seven sample BCDs in H08 are listed in the lower part of Table \[tab:flux\].
-------------- ---------------------- ---------------------- ----------------------- -----------------------
Name *N60* *WIDE-S* *WIDE-L* *N160*
($\lambda =65~\mu$m) ($\lambda =90~\mu$m) ($\lambda =140~\mu$m) ($\lambda =160~\mu$m)
UM 420 $0.46\pm 0.05$ Jy $0.37\pm 0.04$ Jy $0.14\pm 0.03$ Jy $<0.36$ Jy
Mrk 59 $2.3\pm 0.2$ Jy $2.1\pm 0.2$ Jy $1.4\pm 0.3$ Jy $1.2\pm 0.3$ Jy
Mrk 487 $0.45\pm 0.05$ Jy $0.43\pm 0.04$ Jy $0.21\pm 0.06$ Jy $<0.69$ Jy
SBS 1319+579 $0.38\pm 0.04$ Jy $0.29\pm 0.03$ Jy $0.28\pm 0.06$ Jy $<0.59$ Jy
II Zw 40 $6.9\pm 0.7$ Jy $6.6\pm 0.7$ Jy $3.7\pm 0.7$ Jy $3.4\pm 0.9$ Jy
Mrk 7 $0.83\pm 0.23$ Jy $0.84\pm 0.08$ Jy $1.1\pm 0.2$ Jy $1.0\pm 0.3$ Jy
Mrk 71 $3.7\pm 0.4$ Jy $3.1\pm 0.3$ Jy $2.2\pm 0.4$ Jy $<3.0$ Jy
UM 439 $0.42\pm 0.04$ Jy $0.40\pm 0.04$ Jy $0.31\pm 0.06$ Jy $<0.45$ Jy
UM 533 $0.54\pm 0.05$ Jy $0.61\pm 0.06$ Jy $0.54\pm 0.11$ Jy $<0.68$ Jy
II Zw 70 $0.89\pm 0.09$ Jy $0.80\pm 0.24$ Jy $0.75\pm 0.15$ Jy $0.62\pm 0.16$ Jy
II Zw 71 $0.32\pm 0.03$ Jy $0.54\pm 0.05$ Jy $1.3\pm 0.26$ Jy $1.3\pm 0.3$ Jy
-------------- ---------------------- ---------------------- ----------------------- -----------------------
In this paper, we add four BCDs available in the *AKARI* archive[^2]: UM 420, Mrk 59, Mrk 487, and SBS 1319+579 as shown in the upper part of Table \[tab:flux\]. We selected these galaxies among the BCDs observed by FIS after investigating if they were detected at *N60* (65 $\mu$m), *WIDE-S* (90 $\mu$m), and *WIDE-L* (140 $\mu$m). The observing mode (FIS01), the scan speed (8$''$ s$^{-1}$), and the reset interval (2 s) are the same as those adopted for the sample in H08. The method of data reduction and analysis is the same as that in H08, and the measured fluxes are listed in Table \[tab:flux\]. Below we briefly review the data reduction and analysis (see H08 for details).
The raw data were reduced by using the FIS Slow Scan Tool (version 20070914).[^3] Because the detector response is largely affected by a hit of high energy ionizing particle, we used the local flat; that is, we corrected the detector sensitivities by assuming uniformity of the sky brightness [@verdugo07].[^4] The errors caused by the smoothing procedures are well within 10% for *N60* ($65~\mu$m) and *WIDE-S* ($90~\mu$m), 20% for *WIDE-L* ($140~\mu$m), and 25% for *N160* ($160~\mu$m). These error levels are consistent to or somewhat larger than the background fluctuation mainly caused by a hit of high energy ionizing particle, which implies that the background fluctuation is a major component in the errors. The above values are adopted for the errors, but we adopt larger errors for some galaxies which have larger background noises. At *N160* (160 $\mu$m), some objects are not detected. For them, we adopt 3 times the background uncertainty as an upper limit. Since the sample BCDs are compact enough to be treated as point sources with the resolution of FIS, we follow the point-source photometry described in @verdugo07.
Finally [the same colour correction as in H08 was adopted]{} for the *WIDE-L* ($140~\mu$m) fluxes: the colour correction factor was assumed to be 0.93 (the flux was divided by this factor). We did not apply colour correction to the *N60* (65 $\mu$m), *N160* (160 $\mu$m), and *WIDE-S* ($90~\mu$m) fluxes, following H08. For these three bands, the uncertainty caused by not applying colour correction is smaller than the errors put in Table \[tab:flux\].
FIR colours
-----------
The FIR colours are defined by the flux ratio between two bands (Section \[subsec:colour\]). HHS07 investigated the FIR colours of the Milky Way and the Magellanic Clouds by using the Zodi-Subtracted Mission Average (ZSMA) taken by the Diffuse Infrared Background Experiment (DIRBE) of the *Cosmic Background Explorer* (*COBE*). Details of the observational data analysis can be found in @hibi06, who adopted DIRBE bands of 60 $\mu$m, 100 $\mu$m, and 140 $\mu$m and two colours, 60 $\mu$m–100 $\mu$m colour, $(60/100)_\mathrm{cl}$, and 140 $\mu$m–100 $\mu$m colour, $(140/100)_\mathrm{cl}$.
We can also take two colours for our BCD sample. However, the wavelengths of the FIS bands are slightly different from the DIRBE bands. Thus, we apply the following corrections. First, we fit $A\nu^\beta B_\nu (T_\mathrm{d})$ ($A$ is a constant and $\beta$ is the emissivity index) to the 65 $\mu$m and 90 $\mu$m fluxes derive the dust temperature $T_\mathrm{d}$. We denote the dust temperature obtained in this way as $T_\mathrm{d}((65/90)_\mathrm{cl},\,\beta )$. We adopt $\beta =1$ and 2. Then, we estimate the 60 $\mu$m flux by evaluating $A\nu^\beta B_\nu (T_\mathrm{d})$ at $\lambda =60~\mu$m under the values of $A$ and $T_\mathrm{d}((65/90)_\mathrm{cl},\,\beta )$ obtained above. The same procedure is applied for the 140 $\mu$m and 90 $\mu$m fluxes to obtain the 100 $\mu$m flux. In this way, we can obtain $(60/100)_\mathrm{cl}$ and $(140/100)_\mathrm{cl}$ for the BCD sample. In Fig. \[fig:clrdat\], we show the colour–colour relation for the sample.
For comparison, we also show the data of @hibi06 for the Milky Way (the Galactic plane with Galactic latitudes of $|b|<5^\circ$) and the Magellanic Clouds in Fig. \[fig:clrdat\]. @hibi06 found that more than 90% of the data lie on a strong correlation called main correlation, which can be fitted as $$\begin{aligned}
(140/100)_\mathrm{cl}=0.65(60/100)_\mathrm{cl}^{-0.78}
\, .\label{eq:main}\end{aligned}$$ The main correlation also explains the FIR colours of Galactic high latitudes ($|b|>5^\circ$), the LMC, and the SMC (@hibi06 [@hibiphd06]; HHS07). In the Galactic plane, there is another correlation sequence, called subcorrelation in @hibi06: $$\begin{aligned}
(140/100)_\mathrm{cl}=0.93(60/100)_\mathrm{cl}^{-0.56}
\, .\label{eq:sub}\end{aligned}$$ The subcorrelation is not seen in the high Galactic latitudes (@hibiphd06; HHS07). This supports the idea of @hibi06 that the subcorrelation is produced by a contamination of high-ISRF regions, which tend to reside in the Galactic plane.
![Far-infrared colour–colour diagram showing the relation between 60 $\mu$m–100 $\mu$m colour and 140 $\mu$m–100 $\mu$m colour. The filled squares with error bars indicate the current BCD sample, while the crosses and open diamonds show the DIRBE data of the LMC and the SMC, respectively, taken from @hibi06. We also present the Milky Way data (with Galactic latitudes of $|b|<5^\circ$) adopted from @hibi06 by the contours, which show the levels where 50, 80, 90, 95, and 99 per cent of the data are contained. The dashed line and the dotted lines present the fittings for the main correlation (equation \[eq:main\]) and the subcorrelation (equation \[eq:sub\]), respectively. \[fig:clrdat\]](bcd_clr.eps){width="8.5cm"}
It is interesting that not only the LMC and the SMC but also the current BCD sample has consistent FIR colours to the main correlation or the subcorrelation (Fig. \[fig:clrdat\]). This implies that the wavelength dependence of the FIR emissivity is not different among the BCDs, the LMC, the SMC, and the Milky Way. In the following section, we examine if the FIR colours of BCDs can really be reproduced with the emission properties adopted by HHS07, who explained the FIR colours of the Milky Way, the LMC, and the SMC (Section \[sec:model\]).
Results {#sec:result}
=======
Interstellar radiation field {#subsec:dep_isrf}
----------------------------
We present the dependence of FIR colours on ISRF. Here we adopt $A_V=0$ [(i.e. the optically thin extreme for the ISRF; Section \[subsec:colour\])]{} to concentrate only on the effects of ISRF intensity. In Fig. \[fig:clr\_Av0\], we show the FIR colour–colour relation ($(60/100)_\mathrm{cl}$ and $(140/100)_\mathrm{cl}$) calculated for silicate and graphite. We observe that the FIR colours of the BCD sample can be reproduced with $\chi\sim 30$–300 except for the lowest data point (II Zw 71). This is consistent with the conclusion in H08 that the ISRF (especially the UV ISRF, which contributes most to the dust heating) in BCDs can be 100 times higher than the Galactic ISRF in the solar neighbourhood.
![Far-infrared colour–colour relations for silicate and graphite (asterisks and open diamonds, respectively). The different points correspond to $\chi =1$, 3, 10, 30, 100, and 300 from high to low \[140/100\]. [We adopt $A_V=0$ (optically thin extreme) for the ISRF.]{} The same data as shown in Fig. \[fig:clrdat\] as well the the main correlation and subcorrelation (dashed and dotted lines, respectively) are also presented. \[fig:clr\_Av0\]](clr_Av0.eps){width="8.5cm"}
Our results not only confirm the warm *IRAS* (*Infrared Astronomical Satellite*) 60 $\mu$m–100 $\mu$m colours of BCDs [@hoffman89], but also show high temperatures of dust contributing to the emission at $\lambda >100~\mu$m. A few Virgo Cluster BCDs observed by *ISO* (*Infrared Space Observatory*) show relatively cold 170 $\mu$m–100 $\mu$m colours [@popescu02], indicating that the presence of warm dust is not necessarily the case at least in cluster environments.
We also find that the calculated FIR colours cover the area between the main correlation and the subcorrelation. With $\chi\la 10$, the colours trace the main correlation, which is appropriate for the Milky Way, the LMC and the SMC. On the other hand, the FIR colours are rather similar to the subcorrelation for $\chi\ga 100$. While @hibi06 and HHS07 claimed that the subcorrelation can be reproduced with dust emission with multiple temperature components, we see here that the upper part of the subcorrelation can be explained with a single temperature.
Radiative transfer effects {#subsec:radtr}
--------------------------
[If radiative transfer effects are considered,]{} FIR emission comes from dust at different optical depths, where the ISRF intensities are different because of dust extinction. As modeled in Section \[subsec:colour\], the FIR intensity is the sum of all contributions from different optical depths. @hibi06 expect that radiation transfer effects shift the FIR colours toward the subcorrelation on the colour–colour diagram.
In Fig. \[fig:clr\_varAv\], we show the FIR colours for $\chi =3$, 30, and 300 with various values of $A_V$ ($A_V=0$, 0.1, 0.2, 0.5, 1, and 2)[^5]. The FIR colours change little for $A_V\ga 2$, since the emission is negligible from such large optical depths in the wavelength range of interest because of low dust temperatures. As expected by @hibi06, there is a slight trend that the colours move toward the subcorrelation for $\chi\la 30$. However, we observe that the radiative transfer effects shift the FIR colours along the main correlation or the subcorrelation for $\chi\ga 30$. This indicates that the radiative transfer effects are hard to be distinguished from the change of $\chi$ on the colour–colour diagram.
{width="8.5cm"} {width="8.5cm"}
How can we distinguish the variation of $A_V$ from that of $\chi$? We expect that if the change of $A_V$ controls the colour variation there may be a correlation between dust content and FIR colours. Below we further discuss the observed FIR colours in BCDs in terms of the dust content.
Discussion {#sec:discussion}
==========
Dust content and FIR colours {#subsec:dust_content}
----------------------------
The dust mass of the current BCD sample is estimated by the *WIDE-S* ($90~\mu$m) and *WIDE-L* ($140~\mu$m) fluxes, which are suitable to trace the total mass of large grains which have the dominant contribution to the total dust content [@galliano05].[^6] The dust mass $M_\mathrm{d}$ is related to the flux as (e.g. @hildebrand83; H08) $$\begin{aligned}
M_\mathrm{d}=
\frac{F_\nu (\lambda )D^2}{\kappa_\nu B_\nu (T_\mathrm{d})}
\, ,\label{eq:mdust}\end{aligned}$$ where $\kappa_\nu$ is the mass absorption coefficient of dust grains. The mass absorption coefficient is related with the absorption efficiency $Q_\nu$ as (e.g. H08) $$\begin{aligned}
\kappa_\nu =\frac{3Q_\nu}{4as}\, ,\end{aligned}$$ where $s$ is the grain material density. If the grain radius is much smaller than the wavelength, $Q_\nu /a$ is independent of $a$ [@hildebrand83]. The values adopted in Section \[subsec:properties\] indicate $Q_\nu /a=1.14\times 10^{-2}~\mu\mathrm{m}^{-1}$ for silicate and $1.23\times 10^{-2}~\mu\mathrm{m}^{-1}$ for graphite at $\lambda =125~\mu$m. By adopting typical densities for silicate and graphite as $s=3.3$ g cm$^{-3}$ and 2.25 g cm$^{-3}$, respectively, we obtain the mass absorption coefficients at $\lambda =125~\mu$m as $\kappa_{125}=25.9$ and 41.0 cm$^2$ g$^{-1}$ for silicate and graphite, respectively. H08 derived $\kappa_{125}=18.8$ cm$^2$ g$^{-1}$ based on @hildebrand83.
Following H08, we adopt the following expression for $\kappa_\nu$ to estimate the dust mass from the *AKARI* data: $$\begin{aligned}
\kappa_\nu =\kappa_{125}\left(
\frac{\lambda}{125~\mu\mathrm{m}}\right)^{-\beta}\, .\end{aligned}$$ The absorption coefficient adopted in Section \[subsec:properties\] cannot be fitted with a single power law, but the index lies between $\beta =1$ and 2. Thus, if we adopt the wavelength dependence described in Section \[subsec:properties\], the estimated dust mass should be between the dust masses obtained with $\beta =1$ and $\beta =2$. We prefer to adopt this simple power-law expression for $\kappa_\nu$ here, since it can be directly compared with the results in H08 and the simple expression is useful for observational usage.
In Table \[tab:mdust\], we list the dust mass estimated from equation (\[eq:mdust\]). We determine the dust temperature $T_{\rm d}$ by fitting $A\nu^{\beta}B_\nu (T_{\rm d})$ ($A$ is a constant) to the 90 $\mu$m and 140 $\mu$m fluxes. Note that $A$ can be eliminated if we take the flux ratio of those two bands. The estimated dust temperatures are listed in Table \[tab:mdust\], where $T_\mathrm{d}(\beta =1)$ and $T_\mathrm{d}(\beta =2)$ denote the dust temperatures evaluated with $\beta =1$ and 2, respectively. For the mass absorption coefficient, we adopt $\kappa_{125}=25.9$ cm$^2$ g$^{-1}$ (i.e. the value for silicate), but the dust mass is simply proportional to $\kappa_{125}^{-1}$ if another value of $\kappa_{125}$ is adopted. The distance is taken from H08 for the common sample, and is estimated by using the Galactocentric velocity taken from the NASA/IPAC Extragalactic Database (NED)[^7] by assuming a Hubble constant of 75 km s$^{-1}$ Mpc$^{-1}$. The dust masses estimated with $\beta =1$ and 2 are denoted as $M_\mathrm{d}(\beta =1)$ and $M_\mathrm{d}(\beta =2)$, respectively. Although the complex dependence of quantities in equation (\[eq:mdust\]) makes it hard to obtain an analytical estimate of the uncertainty in $M_\mathrm{d}$, we have confirmed that the uncertainty is around a factor of 2 by varying the measured fluxes within the errors. From Table \[tab:mdust\], we observe that $M_\mathrm{d}(\beta =1)$ is systematically smaller than $M_\mathrm{d}(\beta =2)$ by a factor of $\sim 2.5$, mainly because of higher dust temperature for $\beta =1$.
-------------- ------------------ --------------------------------------- --------------------------------------- --------------------------------------- --------------------------------------- --------------------------------- ---------------------------------------
Name $D\,^\mathrm{a}$ $T_\mathrm{d}(\beta =1)\,^\mathrm{b}$ $T_\mathrm{d}(\beta =2)\,^\mathrm{b}$ $M_\mathrm{d}(\beta =1)\,^\mathrm{c}$ $M_\mathrm{d}(\beta =2)\,^\mathrm{c}$ $M_\mathrm{H\, I}\,^\mathrm{d}$ $12+\log \mathrm{(O/H)}\,^\mathrm{e}$
\[Mpc\] \[K\] \[K\] \[$M_\odot$\] \[$M_\odot$\] \[$M_\odot$\]
UM 420 234 ${97^{+249}_{-31}}$ $50^{+20}_{-9}$ $2.5\times 10^5$ $8.7\times 10^5$ — $7.93\pm 0.05$
Mrk 59 10.9 $44^{+15}_{-7}$ $32^{+7}_{-3}$ $2.2\times 10^4$ $6.2\times 10^4$ — $7.99\pm 0.01$
Mrk 487 11.0 $62^{+72}_{-16}$ $40^{+16}_{-7}$ $1.7\times 10^3$ $5.0\times 10^3$ $8.0\times 10^7$ $8.06\pm 0.04$
SBS 1319+579 29.0 $34^{+7}_{-4}$ $26^{+4}_{-2}$ $6.8\times 10^4$ $1.8\times 10^5$ $1.7\times 10^9$ $8.10\pm 0.01$
II Zw 40 9.2 $53^{+20}_{-9}$ $36^{+8}_{-4}$ $2.8\times 10^4$ $8.3\times 10^4$ $2.0\times 10^8$ $8.15\pm 0.02$
Mrk 7 42.4 $28^{+4}_{-2}$ $23^{+3}_{-1}$ $1.0\times 10^6$ $2.7\times 10^6$ $3.6\times 10^9$ $8.54\pm 0.04$
Mrk 71 3.4 $42^{+10}_{-5}$ $31^{+5}_{-3}$ $3.9\times 10^3$ $1.1\times 10^4$ $1.2\times 10^9$ $7.83\pm 0.02$
UM 439 13.1 $39^{+10}_{-5}$ $30^{+4}_{-3}$ $9.8\times 10^3$ $2.7\times 10^4$ $1.7\times 10^8$ $7.98\pm 0.03$
UM 533 10.9 $36^{+7}_{-5}$ $28^{+4}_{-3}$ $1.6\times 10^4$ $4.2\times 10^4$ $5.8\times 10^7$ $8.10\pm 0.04$
II Zw 70 17.0 $34^{+15}_{-5}$ $27^{+8}_{-4}$ $5.9\times 10^4$ $1.5\times 10^5$ $3.6\times 10^8$ $8.11\pm 0.04$
II Zw 71 18.5 $22^{+2}_{-2}$ $19^{+1}_{-2}$ $7.4\times 10^5$ $1.8\times 10^6$ $7.3\times 10^8$ $8.24\pm 0.04$
-------------- ------------------ --------------------------------------- --------------------------------------- --------------------------------------- --------------------------------------- --------------------------------- ---------------------------------------
[ $^\mathrm{a}$ Distance estimated by using the Galactocentric velocity taken fro NED by assuming a Hubble constant of 75 km s$^{-1}$ Mpc$^{-1}$ except for Mrk 71, whose distance is taken from @tolstoy95.\
$^\mathrm{b}$ Dust temperature estimated from the 140 $\mu$m–100 $\mu$m colour (Section \[subsec:dust\_content\]). The emissivity index adopted is also indicated ($\beta =1$ or 2).\
$^\mathrm{c}$ Dust mass estimated in Section \[subsec:dust\_content\]. The emissivity index adopted is also indicated ($\beta =1$ or 2).\
$^\mathrm{d}$ H<span style="font-variant:small-caps;">i</span> gas mass estimated from H<span style="font-variant:small-caps;">i</span> 21 cm line observations. For the latter seven BCDs, the data are compiled in [@hirashita08]. The data for Mrk 487 and SBS 1319+579 are obtained from @hopkins02 and @huchtmeier07, respectively.\
$^\mathrm{e}$ Oxygen abundance compiled in @hopkins02 and @hirashita02 (and the references therein). When the error estimate is not available, we put a typical uncertainty in the measurement (0.04) [@shi05]. ]{}
As an indicator of dust abundance, we take dust-to-gas ratio. For the gas mass, we adopt H<span style="font-variant:small-caps;">i</span> gas mass $M_\mathrm{H\, I}$ estimated from H<span style="font-variant:small-caps;">i</span> 21 cm emission observations. The data of $M_\mathrm{H\, I}$ are already compiled by H08 for the common sample (i.e. the latter seven BCDs in Table \[tab:mdust\]). Among the newly added sample, there is no available H<span style="font-variant:small-caps;">i</span> data for UM 420 and Mrk 59, while the data for Mrk 487 and SBS 1319+579 are obtained from @hopkins02 and @huchtmeier07, respectively (for the latter, we used the formula in @lisenfeld98 to convert the H<span style="font-variant:small-caps;">i</span> flux to $M_\mathrm{H\, I}$). The dust-to-gas ratio $\mathcal{D}$ is defined as $\mathcal{D}\equiv M_\mathrm{d}/M_\mathrm{H\, I}$. We adopt $M_\mathrm{d}(\beta =1)$ for the dust mass, but the following discussion does not change if we adopt $M_\mathrm{d}(\beta =2)$.
Fig. \[fig:dg\_clr\] shows the relations between FIR colour ($(60/100)_\mathrm{cl}$ or $(140/100)_\mathrm{cl}$) and dust-to-gas ratio. We observe correlations with correlation coefficients $r=-0.65$ for the $\log (60/100)_\mathrm{cl}$–$\log\mathcal{D}$ relation and $r=0.89$ for the $\log (140/100)_\mathrm{cl}$–$\log\mathcal{D}$ relation. These correlations indicate that the dust temperature tends to be high in dust-poor objects. We present not only the sample BCDs but also the data of the Milky Way, the LMC and the SMC. The dust-to-gas ratio of the Milky Way is assumed to be 0.006 [@spitzer78], while the dust-to-gas ratios of the LMC and the SMC are assumed to be 1/3 and 1/5 times of the value of the Milky Way [@pei92]. For the FIR colours, we adopt the peak of the contours in Fig. \[fig:clrdat\] for the Milky Way, and the average of the data points for the LMC and the SMC. Those three points also follow the trend found by the BCDs.
{width="8.5cm"} {width="8.5cm"}
A possible interpretation of the correlation in Fig. \[fig:dg\_clr\] is that the relatively low dust temperatures in dust-rich BCDs result from the shielding (extinction) of stellar radiation. In order to examine this possibility, we calculate the relation between FIR colours and $A_V$ by using the models (same as Fig. \[fig:clr\_varAv\]). The results are shown in Fig. \[fig:Av\_clr\], where we find that both the variation of $(60/100)_\mathrm{cl}$ and that of $(140/100)_\mathrm{cl}$ with a single value of $\chi$ are too small to explain the observed large diversity in these colours in Fig. \[fig:dg\_clr\]. The colours are not so sensitive to the change of $A_V$ partly because the strongest contribution from $A_V\sim 0$, where the dust temperature is the highest, is always present. In particular, if $A_V\ga 2$, the colour is insensitive to $A_V$ since the dust temperature at such deep optical depths is too low to contribute to the total emission. Thus, the variation of the FIR colours cannot be reproduced only with a variation of dust optical depth, but rather it should reflect the correlation (i) between dust-to-gas ratio and dust heating itself (i.e., ISRF), or (ii) between dust-to-gas ratio and dust properties (i.e., absorption coefficient and/or grain size distribution).
{width="8.5cm"} {width="8.5cm"} {width="8.5cm"} {width="8.5cm"}
It is difficult to survey all the possible cases for (ii), since there are few, if any, constraints on the dust properties in BCDs. However, we can discuss some cases for (ii) as follows. HHS07 show that $(140/100)_\mathrm{cl}$ changes only slightly along with the change of grain size distribution, since it reflects the equilibrium grain temperature. Thus, it is hard to reproduce the clear correlation between $(140/100)_\mathrm{cl}$ and dust-to-gas ratio only with the change of grain size distribution. The change of grain material as a function of dust-to-gas ratio may change the grain absorption coefficient, leading to variation of dust temperature as a function of dust-to-gas ratio. However, we have shown that the colour–colour relation of the BCDs can be explained by the emissivity that also explains the colour–colour relation of the Milky Way. In this context, it is not probable that the dust property is the main driver for the change of FIR colours.
The possibility (i) is interesting to investigate, since the correlation in Fig. \[fig:dg\_clr\] is a natural extension of the sequence of the Milky Way, the LMC, and the SMC. For those “nearby” three galaxies, it is known that the typical ISRFs are different [e.g. @welty06], and indeed the FIR colour sequence can be explained by the variation of the ISRF (@hibi06, HHS07). Thus, in the following subsection, we discuss a possibility that the ISRF changes as a function of dust-to-gas ratio.
Correlation between dust-to-gas ratio and ISRF?
-----------------------------------------------
Here we interpret the correlation between dust-to-gas ratio and FIR colours as the correlation between dust-to-gas ratio and ISRF. Since the main heating source of dust grains is UV photons coming from the young stellar population [@buat96], the dust-heating ISRF reflects the spatial concentration of recent star formation activity. Thus, the correlation between dust-to-gas ratio and ISRF indicates that dust-poor galaxies tend to host concentrated star-forming regions.
There are some theoretical models which suggest that dust plays an active role in determining the properties of star formation. Indeed @hirashita04 show that the shielding of UV heating photons plays an important role in determining the strength of star formation: If the optical depth of dust becomes large, the UV heating photons are efficiently blocked and the gas efficiently cools to produce a favourable condition for the star formation. Thus, the optical depth of dust in UV is important, and a burst of star formation could become possible if the UV optical depth exceeds one. In this scenario, the condition for star formation can be written as $$\begin{aligned}
\kappa\mathcal{D}\rho R\sim 1\, ,\label{eq:tau1}\end{aligned}$$ where $\kappa$ is the UV absorption coefficient per dust mass and $R$ is the size of the star-forming region. On the other hand, we assume that a constant fraction of gas mass is converted to stars: $L\propto M_\mathrm{gas}$, where $M_\mathrm{gas}(\sim\rho R^3)$ is the gas mass, and $L$ is the stellar UV luminosity. Since $\chi\propto L/R^2\propto M_\mathrm{gas}/R^2\sim
\rho R$, we obtain $\chi\propto\mathcal{D}^{-1}$ by using equation (\[eq:tau1\]). Normalizing the quantities to the Milky Way values (i.e. $\chi =1$ and $\mathcal{D}=0.006$), we obtain $$\begin{aligned}
\chi =\left(\frac{\mathcal{D}}{0.006}\right)^{-1}\, .
\label{eq:chi_dg}\end{aligned}$$
In Fig. \[fig:dg\_clr\], we calculate the relation between the FIR colours and $\mathcal{D}$ by assuming equation (\[eq:chi\_dg\]). [The optically thin extreme (i.e. $A_V=0$) for the ISRF is adopted to focus on the effect of $\chi$.]{} The results indeed traces the trend of the observational data. Thus, the above argument on dust shielding of UV photons (equation \[eq:tau1\]) as the condition for a burst of star formation is compatible with the data.
Equation (\[eq:tau1\]) implies that the size-density relation of star-forming region becomes roughly a relation with constant column density. Indeed, the observational data presented by @hirashita09 support constant column density of star-forming regions in BCDs (i.e. $\rho\propto R^{-1}$). @pak98 also find that the column density of gas is inversely proportional to the dust-to-gas ratio based on an analysis of photodissociation regions in the Milky Way, the LMC, and the SMC, which is consistent with our picture. @mckee89 show that an equilibrium between the gravitational contraction and the energy injection from stars is achieved at $A_V\sim 4$–8 based on the stability argument on photoionization-regulated star formation, in which shielding of photoionizing photons permits ambipolar diffusion to proceed and stars to form. Although the physical processes considered in @mckee89 are more detailed and sophisticated, the necessity of shielding is common with our picture expressed in equation (\[eq:tau1\]).
FIR colour and metallicity {#subsec:metal}
--------------------------
The observational correlation between FIR colour and dust-to-gas ratio shown in this paper has confirmed the results of @engelbracht08, who show the correlation between *Spitzer* MIPS 70 $\mu$m–160 $\mu$m colour and metallicity. Since there is a correlation between dust-to-gas ratio and metallicity as shown by them and also by H08, the correlation with metallicity is equivalent to the correlation with dust-to-gas ratio. For the detailed discussions on the relation between dust-to-gas ratio and metallicity in BCDs, see H08 and @lisenfeld98.
Fig. \[fig:metal\_clr\] shows the relation between FIR colour and metallicity (oxygen abundance in gas phase) for the BCD sample. The data of oxygen abundance are compiled in @hirashita02 and @hopkins02, and summarized in Table \[tab:mdust\]. We observe correlations with $r=-0.45$ for the relation between $\log (60/100)_\mathrm{cl}$ and $12+\log (\mathrm{O/H})$ and $r=0.53$ for the relation between $\log (140/100)_\mathrm{cl}$ and $12+\log (\mathrm{O/H})$. We also show the data points for the Milky Way, the LMC, and the SMC. The Milky Way gas metallicity is represented by the Orion Nebula ($12+\log (\mathrm{O/H})=8.58$), which has a similar metallicity to local B-type stars [@peimbert87; @gies92; @kilian92; @mathis00]. For the LMC and the SMC, we adopt the oxygen abundance from @dufour84 ($12+\log (\mathrm{O/H})=8.43$ and 8.02, respectively). We confirm that metal-poor galaxies tend to have high dust temperatures.
{width="8.5cm"} {width="8.5cm"}
We should keep in mind that the story of star formation properties in terms of metallicity is not so simple as we discussed above. @engelbracht08 show that the dust temperature shows the peak around $12+\log (\mathrm{O/H})\sim 8$. Due to our small sample in such a low metallicity range, we cannot confirm it, but their results imply that the size and density of star-forming region is not a monotonic function of metallicity. There are also pieces of evidence that there are two categories of BCDs: “active” and “passive”. The former hosts compact and intense star formation activities, while the latter has diffuse star formation properties [@hunt02; @hirashita04]. Even with a similar metallicity, such a dichotomy exists, indicating that the star formation properties are not determined simply by the metallicity. Active BCDs tend to be luminous in FIR because of large dust optical depth [@hirashita04]. Thus, it would be possible that the current sample is biased to the “active” class. The presence of BCDs hosting cold dust in Virgo Cluster [@popescu02] also implies that such a bias could be present. This kind of bias should be examined in the future with more sensitive facilities.
Finally, it is worth questioning which of dust-to-gas ratio and metallicity is more fundamental in regulating dust temperature. Larger absolute values of correlation coefficient found in Fig. \[fig:dg\_clr\] than those in Fig. \[fig:metal\_clr\] imply that dust-to-gas ratio rather than metallicity is more strongly connected with dust temperature. Thus, we here suggest that dust-to-gas ratio is more fundamental than metallicity in regulating the dust temperature. This should be further examined with a larger sample.
Importance of submillimetre data {#subsec:submm}
--------------------------------
It is expected that the radiative transfer effect, i.e.the shielding effect of dust can be seen more clearly in submillimetre (submm) than in FIR, since the submm emission can trace dust with lower temperature [e.g. @galliano03]. Although the submm data of the current BCD sample are still lacking except for those of IIZw40, future more sensitive submm facilities such as *Herschel* and ALMA will increase the data.
We adopt $\lambda =850~\mu$m as a representative submm wavelength, and examine the relation between $(850/100)_\mathrm{cl}$ and $(60/100)_\mathrm{cl}$; that is, we adopt $\lambda =850~\mu$m instead of $\lambda =140~\mu$m, which is adopted in the rest of this paper. In Fig. \[fig:clr\_Av0\_submm\], we present the colour–colour relation with $A_V=0$ (i.e. without the radiative transfer effect; [Section \[subsec:colour\]]{}). We also show the nearby galaxy sample observed at $\lambda =850~\mu$m by @dunne00, who also list the *IRAS* 60 $\mu$m and 100 $\mu$m fluxes. We observe that the prediction traces the trend of the data, indicating that the submm flux is naturally explained by the natural extension of the FIR SED. This means that a very cold component contributing only to the submm flux is not necessary for a major part of the nearby galaxies.
@galliano03 [@galliano05] argue that a very cold component should be introduced for metal-poor dwarf galaxies. In particular, IIZw40, which is included also in the current sample, shows a clear excess of the submm flux. In Fig. \[fig:clr\_Av0\_submm\], we also plot the data point taken from @galliano05 for IIZw40 (see also @hunt05). Indeed, we observe a clear deviation from the model prediction, which can be interpreted as a significant contribution from a very cold component. They also propose that such very cold dust should be located in an environment where the stellar radiation is strongly shielded. In order to examine the effect of shielding (i.e.radiative transfer), we show the results with $A_V>0$ in Fig. \[fig:clr\_varAv\_submm\]. We observe that $(850/100)_\mathrm{cl}$ increases even for $A_V>2$ while $(60/100)_\mathrm{cl}$ is almost constant for such large $A_V$. This confirms that 850 $\mu$m flux is also sensitive to the shielded cold dust grains. Fig. \[fig:clr\_varAv\_submm\] also demonstrates that the data point of IIZw40 is explained with $A_V\sim 5$ with $\chi\sim 300$. Thus we conclude the existence of shielded cold dust in II Zw 40 (and in other galaxies whose data points deviate rightwards in Fig. \[fig:clr\_Av0\_submm\]). @galliano05 also show that the UV ISRF of II Zw 40 is higher than that of the Milky Way by two orders of magnitude.
![$(60/100)_\mathrm{cl}$–$(850/100)_\mathrm{cl}$ relations for silicate and graphite (asterisks and open diamonds, respectively). The different points correspond to $\chi =3$, 10, 30, 100, and 300 from high to low $(850/100)_\mathrm{cl}$. [The optically thin extreme for the ISRF ($A_V=0$) is adopted.]{} The filled squares indicate the observational data of nearby galaxies in @dunne00 and the open triangle shows the data point of II Zw 40 taken from @galliano05. \[fig:clr\_Av0\_submm\]](clr_Av0_submm.eps){width="8.5cm"}
{width="8.5cm"} {width="8.5cm"}
Although our interpretation of the submm excess as a contribution from the shielded cold dust component is in agreement with the analysis by @galliano05, there is quantitative difference between our results and theirs. @galliano05 adopted $Q_\mathrm{abs}\propto\lambda^{-2}$ at $\lambda >100~\mu$m. If the absorption efficiency is normalized to the value at $\lambda =100~\mu$m, the absorption efficiency adopted in this paper (equation \[eq:reach\]) is 2 times larger than that used in @galliano05 at $\lambda =850~\mu$m. Therefore, we require less very cold dust component than @galliano05 by a factor of 2.
Common dust optical properties?
-------------------------------
In this paper, we have adopted a common FIR dust optical properties; that is, we have applied the same absorption efficiency $Q_\mathrm{abs}$ and grain size distribution as adopted in our previous model for the Milky Way and the Magellanic Clouds (HHS07). The same dust optical properties are also consistent with the FIR colours of BCDs. [@hibi06] and HHS07 also argue that those dust optical properties also provide a good statistical fit to the FIR colours of nearby galaxies. Therefore, we suggest that the absorption efficiency in the form of equation (\[eq:reach\]), which @reach95 adopted to fit the dust emission spectra in the Milky Way, is generally applicable in galactic environments. As shown above, the submm emission from nearby galaxies can also be explained by the same dust emissivity.
It is also important to stress that the classical dust emissivity model by @draine84 cannot explain the FIR colour–colour relation of the nearby galaxies (@hibi06; HHS07). As indicated by the emissivity assumed by @reach95, it is better to adopt $\beta\sim 1$ for $100~\mu\mathrm{m}\la\lambda\la\lambda_1\sim 200~\mu$m rather than to assume $\beta\sim 2$ in the entire FIR wavelength range longer than 100 $\mu$m. Some amorphous materials indeed show $\beta <2$ [e.g. @agladze96]. Amorphous grains are expected to form by the irradiation of cosmic rays [@jager03].
Theoretically the break of the dust emissivity at $\lambda_1$ might be associated with the energy splitting of the ground state because of irregular amorphous structure. @meny07 show that the absorption coefficient of amorphous dust has a break at $\lambda\sim\lambda_\mathrm{m}$, which corresponds to the cut-off energy $\hbar\omega_\mathrm{m}$ of the energy splitting due to the amorphous structure ($\omega_\mathrm{m}\equiv 2\pi c/\lambda_\mathrm{m}$). Therefore, a possible interpretation is that $\lambda_\mathrm{m}\sim\lambda_1$. Since @meny07 suggest $\lambda_\mathrm{m}\sim 700~\mu$m based on experimental data, the cut-off energy corresponding to $\lambda_\mathrm{m}\sim 200~\mu$m may be too high. Nevertheless it is still worth investigating amorphous materials with higher cut-off energy as a candidate of cosmic dust.
Conclusion {#sec:conclusion}
==========
We have investigated the properties of FIR emission of a sample of BCDs observed by *AKARI*, especially focusing on FIR colours and dust temperature. We have utilized the data at $\lambda =65~\mu$m, 90 $\mu$m, and 140 $\mu$m, and have examined the relation between 60 $\mu$m–100 $\mu$m colour, $(60/100)_\mathrm{cl}$, and 140 $\mu$m–100 $\mu$m colour $(140/100)_\mathrm{cl}$. Then, we have found that the FIR colours of the BCDs are located at a natural high-temperature extension of the DIRBE data of the Milky Way, the LMC and the SMC on the colour–colour diagram. We have explained the FIR colours also theoretically by assuming the same absorption efficiency, which may be appropriate for amorphous dust grains, and the same grain size distribution as the Milky Way dust. We have also shown that it is not easy to distinguish between a large dust optical depth and a low dust temperature only with FIR colours although addition of submillimetre data relax this degeneracy.
In order to examine if the dust optical depth plays an important role in determining the dust temperature, we have investigated the correlation between FIR colour (dust temperature) and dust-to-gas ratio. We have found that the dust temperature tends to become high as the dust-to-gas ratio decreases as would be expected from the shielding effect of stellar radiation by dust. However, we fail to explain this trend quantitatively only by the effect of dust optical depth. Thus, we conclude that there is a relation between dust-to-gas ratio and interstellar radiation field (ISRF) intensity. This relation is consistent with a “constant dust optical depth” of star-forming regions in BCDs, implying that dust extinction plays an important role in determining the condition for a burst of star formation.
The correlation between dust-to-gas ratio and dust temperature is equivalent to that between metallicity and dust temperature found in @engelbracht08, since there is a correlation between dust-to-gas ratio and metallicity. By comparing the correlation strengths, we propose that dust-to-gas ratio is more fundamental than metallicity in regulating dust temperature, although we will have to confirm this with a larger sample.
Acknowledgments {#acknowledgments .unnumbered}
===============
We thank the anonymous referee for useful comments which improved this paper considerably. We are grateful to T. Onaka, H. Kaneda, and Y. Hibi for helpful discussions. We thank all members of *AKARI* project for their continuous help and support. This research has made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.
[99]{} Agladze, N. I., Sievers, A. J., Jones, S. A., Burlitch, J. M., & Beckwith, S. V. W. 1996, ApJ, 462, 1026 Bianchi, S., Davies, J. I., & Alton, P. B. 1999, A&A, 344, L1 Bohren, C. F., & Huffman, D. R. 1983, Absorption and Scattering of Light by Small Particles, Wiley, New York Boselli, A., Gavazzi, G., & Sanvito, G. 2003, A&A, 402, 37 Buat, V., & Xu, C. 1996, A&A, 306, 61 Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245 Désert, F.-X., Boulanger, F., & Puget, J. L. 1990, A&A, 237, 215 Dopita, M. A., et al. 2005, ApJ, 619, 755 Draine, B. T., & Lee, H. M. 1984, ApJ, 285, 89 Draine, B. T., & Li, A. 2001, ApJ, 551, 807 Dufour, R. J. 1984, in Structure and Evolution of the Magellanic Clouds, ed. S. van den Bergh, & K. S. de Boer (Kluwer, Dordrecht), p. 353 Dunne, L., Eales, S., Edmunds, M., Ivison, R., Alexander, P., Clements, D. L. 2000, MNRAS, 315, 115 Dwek, E., et al. 1997, ApJ, 475, 565 Engelbracht, C. W., Rieke, G. H., Gordon, K. D., Smith, J.-D. T., Werner, M. W., Moustakas, J., Willmer, C. N. A., & Vanzi, L. 2008, ApJ, 678, 804 Galliano, F., Madden, S. C., Jones, A. P., Wilson, C. D., & Bernard, J.-P. 2007, A&A, 434, 867 Galliano, F., Madden, S. C., Jones, A. P., Wilson, C. D., Bernard, J.-P., & Le Peintre, F. 2003, A&A, 407, 159 Gies, D. R., & Lambert, D. L. 1992, ApJ, 387, 673 Hibi, Y. 2006, PhD Thesis, Nagoya University Hibi, Y., Shibai, H., Kawada, M., Ootsubo, T., & Hirashita, H. 2006, PASJ, 58, 509 Hildebrand, R. H. 1983, QJRAS, 24, 267 Hirashita, H., Tajiri, Y. Y., & Kamaya, H. 2002, A&A, 388, 439 Hirashita, H., Hibi, Y., & Shibai, H. 2007, MNRAS, 379, 974 (HHS07) Hirashita, H., & Hunt, L. K. 2004, A&A, 460, 67 Hirashita, H., & Hunt, L. K. 2009, A&A, submitted Hirashita, H., Kaneda, H., Onaka, T., & Suzuki, T. 2008, PASJ, 60, S477 (H08) Hoffman, G. L., Helou, G., Salpeter, E. E., & Lewis, B. M. 1989, ApJ, 339, 812 Hopkins, A. M., Schulte-Ladbeck, R. E., & Drozdovsky, I. O. 2002, AJ, 124, 862 Huchtmeier, W. K., Petrosian, A., Gopal-Krishna, Kunth, D. 2007, A&A, 462, 919 Hunt, L. K., Giovanardi, C., & Helou, G. 2002, A&A, 394, 873 Hunt, L. K., Bianchi, S., & Maiolino, R. 2005, A&A, 434, 849 Iglesias-Páramo, J., Buat, V., Donas, J., Boselli, A., & Milliard, B. 2004, A&A, 419, 109 Inoue, A. K., Hirashita, H., & Kamaya, H. 2000, PASJ, 52, 539 Jäger, C., Fabian, D., Schrempel, F., Dorschner, J., Henning, Th., & Wesch, W. 2003, A&A, 401, 57 Kawada, M., et al. 2007, PASJ, 59, S389 Kennicutt, R. C., Jr. 1998, ARA&A, 36, 189 Kilian, J. 1992, A&A, 262, 171 Li, A., & Draine, B. T. 2001, ApJ, 554, 778 Lisenfeld, U., & Ferrara, A. 1998, ApJ, 496, 145 Madden, S. C., Galliano, F., Jones, A. P., & Sauvage, M. 2006, A&A, 446, 877 Maiolino, R., Schneider, R., Oliva, E., Bianchi, S., Ferrara, A., Mannucci, F., Pedani, M., Roca Sogorb, M. 2004, Nature, 431, 533 Mathis, J. S. 2000, in Allen’s Astrophysical Quantities, Fourth Edition, ed. A. N. Cox, (Springer, New York), p. 523 Mathis, J. S., Mezger, P. G., & Panagia, N. 1983, A&A, 128, 212 Mathis, J. S., Rumpl, W., & Nordsieck, K. H. 1977, ApJ, 217, 425 McKee, C. F. 1989, ApJ, 345, 782 Meny, C., Gromov, V., Boudet, N., Bernard, J.-Ph., Paradis, D., & Nayral, C. 2007, A&A, 468, 171 Murakami, H., et al. 2007, PASJ, 59, S369 Nagata, H., Shibai, H., Takeuchi, T. T., & Onaka, T. 2002, PASJ, 54, 695 Nozawa, T., Kozasa, T., Umeda, H., Maeda, K., & Nomoto, K. 2003, ApJ, 598, 785 Onaka, T., Tokura, D., Sakon, I., Tajiri, Y. Y., Takagi, T., & Shibai, H. 2007, ApJ, 654, 844 Pak, S., Jaffe, D. T., van Dishoeck, E. F., Johansson, L. E. B., & Booth, R. S. 1998, ApJ, 498, 735 Pei, Y. C., 1992, ApJ, 395, 130 Peimbert, M. 1987, in Star Forming Regions, ed. M. Peimbert & J. Jugaku (Reidel, Dordrecht), p. 111 Popescu, C. C., Tuffs, R. J., Völk, H. J., Pierini, D., & Madore, B. F. 2002, ApJ, 567, 221 Reach, W. T., et al. 1995, ApJ, 451, 188 Sakon, I., Onaka, T., Ishihara, D., Ootsubo, T., Yamamura, I., Tanabé, & Roellig, T. L. 2004, ApJ, 609, 203 Sargent, W. L. W., & Searle, L. 1970, ApJ, 162, L155 Schneider, R., Ferrara, A., & Salvaterra, R. 2004, MNRAS, 351, 1379 Shi, F., Kong, X., Li, C., & Cheng, F. Z. 2005, A&A, 437, 849 Spitzer, L., Jr 1978, Physical Processes in the Interstellar Medium, New York, Wiley Takagi, T., Vansevičius, V., & Arimoto, N. 2003, PASJ, 55, 385 Takeuchi, T. T., Ishii, T. T., Nozawa, T., Kozasa, T., & Hirashita, H. 2005, MNRAS, 362, 592 Thuan, T. X., Sauvage, M., & Madden, S. 1999, ApJ, 516, 783 Todini, P., & Ferrara, A. 2001, MNRAS, 325, 726 Tolstoy, E., Saha, A., Hoessel, J. G., & McQuade, K. 1995, AJ, 110, 1640 van Zee, L., Skillman, E. D., & Salzer, J. J. 1998, AJ, 116, 1186 Verdugo, E., Yamamura, I., & Pearson, C. P. 2007, AKARI FIS Data User Manual Version 1.3 Weingartner, J. C., & Draine, B. T. 2001, ApJ, 548, 296 Welty, D. E., Federman, S. R., Gredel, R., Thorburn, J. A., & Lambert, D. L. 2006, ApJS, 165, 138
[^1]: E-mail: hirashita@asiaa.sinica.edu.tw
[^2]: http://darts.isas.jaxa.jp/astro/akari/.
[^3]: http://www.ir.isas.ac.jp/ASTRO-F/Observation/.
[^4]: We referred to *AKARI* FIS Data User Manual Version 2 for the data reduction and analysis (http://www.ir.isas.ac.jp/ASTRO-F/Observation/). When we ran the pipeline command [ss\_run\_ss]{}, we applied options [/local,/smooth,width\_filter=80]{}: The flat field was built from the observed sky, and boxcar smoothing with a filter width of 80 s in the time series data was applied to remove remaining background offsets among the pixels.
[^5]: [As stressed in Section \[subsec:colour\], the case of $A_V=0$ is considered as the optically thin extreme for the ISRF.]{}
[^6]: @galliano05 also introduce a very cold dust component contributing to the submillimetre flux, but its abundance is sensitive to the assumed emissivity index of large grains (Section \[subsec:submm\]). Thus, we here concentrate on the dust component traced in FIR and do not consider very cold dust.
[^7]: http://nedwww.ipac.caltech.edu/.
|
{
"pile_set_name": "ArXiv"
}
|
Monségur, Lot-et-Garonne
Monségur is a commune in the Lot-et-Garonne department in south-western France.
See also
Communes of the Lot-et-Garonne department
Montsegur
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Q:
Should I also index columns included in a PRIMARY KEY?
This question suddenly popped into my head... I have a table that ties two other tables together based on their ID. The CREATE TABLE looks like this:
CREATE TABLE `ticket_contact` (
`ticket_id` INT NOT NULL,
`entity_id` INT NOT NULL,
`notify` INT NOT NULL DEFAULT 0,
PRIMARY KEY (`ticket_id`, `entity_id`),
KEY `ticket_id` (`ticket_id`),
KEY `entity_id` (`entity_id`)
)
I'm wondering if there is any need to include those last two KEY lines. Will it give me improved speed with the following queries, or will individual columns within a PRIMARY KEY automatically be indexed?
SELECT * FROM ticket_contact WHERE ticket_id=1;
SELECT * FROM ticket_contact WHERE entity_id=1;
A:
An index created by PRIMARY KEY is the same as any other (potentially composite) UNIQUE index. So you needn't create a separate index for ticket_id as it is included as the major column in the (ticket_id, entity_id) index.
You would want to create a separate index over entity_id if you were commonly doing queries using that column independently of the ticket_id.
A:
Think of a PK and an Index as simply a way of forcing your records into sorted order. Databases can search through sorted data much much faster than unsorted data (log n vs linear time).
A composite primary key sorts in the order of columns given, so PK (ticket_id, entity_id) sorts ticket_id ASC, entity_id ASC. Since your PK already sorts ticket_ids, it covers the index on ticket_id.
However, sorting by entity_id ASC with no other columns results in a different sort order. If you needed to query on entity_id often, MySQL will perform an index scan (searching through each ticket_id, then binary_searching the results for the matching entity_id). Your separate index on entity_id will make queries for entity_id much faster.
|
{
"pile_set_name": "StackExchange"
}
|
HER2 as a prognostic factor in breast cancer.
HER2 amplification/overexpression is a marker of poor prognosis in breast cancer. The prognostic impact of HER2 positivity is lower in node-negative compared with node-positive women. The only significant, independent prognostic factors in breast cancer are node status, HER2 status and menopausal status. HER2-positive tumors also contain p53 abnormalities, tend to be hormone receptor and bcl-2 negative, have lymphoid infiltration (LI) and a high mitotic index. Patients with LI who are HER2 positive have a better prognosis than those who are HER2 negative, whereas HER2-positive patients without LI have a significantly worse prognosis than HER2-negative patients. Morphological and biological alterations appear to identify two categories of breast tumor. Two hypotheses may explain the progression to two tumor types: (1) atypical ductal hyperplasia (ADH) is a precursor of ductal carcinoma in situ (DCIS), which is a precursor of invasive ductal carcinoma (IDC); or (2) ADH is a precursor of HER2-negative IDC whereas DCIS is a precursor of HER2-positive IDC. The second theory fits well with two breast cancer subsets and the characteristics of ADH and DCIS. The first type of IDC occurs in older patients, progresses slowly due to estrogen dependency but is aggressive long term. The other type progresses rapidly, is HER2 positive and is more likely to occur in young patients.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Egy 72 éves májrákos férfin végezték el hazánkban először az úgynevezett radioembolizációs daganatterápiát, közölte a Semmelweis Egyetem. Ez az eljárás a hagyományos onkológiai kezelésre nem reagáló, sebészileg el nem távolítható primer és áttétes májdaganatokban szenvedőknek adhat új reményt. Biztatóak az első eredmények.
A hazánkban most először végzett intervenciós radiológiai daganatterápiát egy 72 éves, jó állapotú páciensnél alkalmazták még december elején. A betegnél a nyelvgyök területéről kiinduló daganat okozott áttétet a májban.
Az eljárás lényege, hogy a kar felől katéterrel bejutva a máj tápláló erein keresztül a tumor vérellátásába mikroszkopikus méretű sugárforrások tízezreit juttatják be. Ezek az apró sugárzó részecskék a daganat ereiben elakadva rendkívül nagy dózisú, de csak a tumorban hatást kifejtő, belső sugárterápiás kezelést jelentenek.
Bánsághi Zoltán, a Radiológiai Klinika igazgatóhelyettese, a nukleáris medicina és képalkotó szakemberből, intervenciós radiológusból és klinikai onkológusból álló munkacsoport vezetője azt mondta: a kezelés fájdalommentes, körülbelül 40 perces érfestésként élhető meg, amely után mindössze 24 órás megfigyelésre van csak szükség.
A speciális kezelésen átesett beteg jól van, már másnap hazaengedték. A kezelés után két hétig még gyenge és fáradékony volt, de a január elején végzett PET CT vizsgálat alapján a korábban 10 centiméternél nagyobb, hagyományos kezelésre nem reagáló daganat térfogatának több mint fele már biztosan elpusztult, mondta Bánsághi Zoltán.
A világon már másfél évtizede használt eljárást olyan elsődleges májrák, valamint más szervek okozta májáttétek esetén alkalmazzák, ahol sebészi ellátás nem kerülhet szóba, más daganatellenes terápiák sem hoztak eredményt, a gyógyszerek okozta súlyos, elviselhetetlen mellékhatások miatt nincs más lehetőség vagy amikor a daganat már nem reagál az addig alkalmazott kezelésre (rezisztencia) .
A társadalombiztosítás által egyedi méltányosság alapján finanszírozott beavatkozást a Semmelweis Egyetem Radiológiai Klinikáján a Nukleáris Medicina Központtal közösen, az egyetem Onkológiai Központjával és Sugárvédelmi Szolgálatával közreműködésben végezték. A radioembolizációs daganatterápia eredményei a nem operálható esetekben is biztatóak. Számos adat arra utal, hogy az eljárással a műtéti ellátásra nem alkalmas esetek egy részénél csökkenteni lehet a daganat méretét és növelni a daganatmentes májterület térfogatát, ami műthetővé teheti a pácienst.
Az első hazai beavatkozást mintegy kétéves előkészítő munka előzte meg. Az úttörő eljárást egyelőre kizárólag a Semmelweis Egyetemen végezhetik, mondta Bérczi Viktor, a Radiológiai Klinika igazgatója. Most tervezik a második hazai kezelést, amely várhatóan január végén megtörténhet.
Magyarországon az elmúlt négy évtizedben mind a máj elsődleges, mind a más szervből kiinduló, de a májba áttétet adó daganatok száma növekedett. Éves előfordulásuk összesen 8-9 ezer közé tehető, és körülbelül 6 ezer esetben a daganatos máj sebészileg operálhatatlan, az esetek felében pedig sikeres műtét után is számítani lehet a daganat kiújulására.
|
{
"pile_set_name": "OpenWebText2"
}
|
Q:
optional multiple oneToMany relation in symfony2
I have entities in Doctrine Symfony2: User, Channel, Video and Comment; user can report one of them. I designed Report entity with these fields:
userId
status
reportTime
description
how can I reference to reported Entity ?? because all reported fields are similar for all entities I want to use just one table for Report and add these fields to Report Entity:
referenceEntityName(a string and may be one of these: User, Channel, Video, Comment)
Channel(ManytoOne relation to Channel entity)
Video(ManytoOne relation to Video entity)
Comment(ManytoOne relation to Comment entity)
User(ManytoOne relation to User entity)
Is this best practice or I should create separate tables for each kind of report ??
Edit:
based on @Alex answer, I improved Report class and add these methods:
setEntity($entity){
if ($obj instanceof Video){
$this->referenceEntityName = 'Video';
$this->setVideo();
}
elseif($obj instanceof Comment){
$this->referenceEntityName == 'Comment'
$this->setComment();
}
//...
}
getEntity(){
if($this->referenceEntityName == 'Video'){
$this->getVideo()
}// ifelse statements for other entities ...
}
I till have 4 relation that just one of them is used for each instance, isn't it a bit messy!?
and again is this best practice or I should do something else?
what if I want to use FormBuilder class, isn't there any problem??
A:
In a simple solution, whereby for example you only had Users (and not Videos, Comments and Channels), the solution would be simple; each User can have many Reports, and each Report must belong to only one User. This is a one-to-many relationship - one User has many Reports. In Symfony 2 and Doctrine, this would be modelled as such:
// src/Acme/DemoBundle/Entity/User.php
// ...
use Doctrine\Common\Collections\ArrayCollection;
class User
{
// ...
/**
* @ORM\OneToMany(targetEntity="Report", mappedBy="user")
*/
protected $reports;
public function __construct()
{
$this->reports = new ArrayCollection();
}
// ...
}
and
// src/Acme/DemoBundle/Entity/Report.php
// ...
class Report
{
// ...
/**
* @ORM\ManyToOne(targetEntity="User", inversedBy="reports")
* @ORM\JoinColumn(name="user_id", referencedColumnName="id")
*/
protected $user;
// ...
}
In this instance, to create a Report and associate it with a User, we would:
// get the User the Report will belong to
$user = $em->getRepository('AcmeDemoBundle:User')->find(1);
// create the Report
$report = new Report();
// add the User to the Report
$report->setUser($user);
// then persist it, etc ...
Note, the setUser() method is available because the console command was run to generate them automatically. This is highly recommended as it created the necessary type hinting for you. For pre Symfony 2.5 installations, the command is:
php app/console doctrine:generate:entities Acme
>= 2.5 installations, the command is:
php bin/console doctrine:generate:entities Acme
Your requirements complicate this simple example somewhat, as Reports can also belong to Comments and Videos etc. For the sake of the example, let's call these things Entities. A bad approach would be to simply add 3 new properties to the Report, one for each of the new Entities, and then add 3 new setter methods for the Entities. This is bad for 2 reasons: a Report will only ever belong to one of the Entities, and therefore 3 of the properties and setter methods will never be used for each Report entity. Secondly, if you add a new Entity to your business model, or remove one, you need to edit your Report entity, and also the database schema.
A better method is to simply have one property and set method in your Report, that can be applied to all of your Entities. So instead of calling setUser, we could call a setEntity, and have it accept any of the 4. With this approach in mind, let's look back at the first example, and take note of the type hinting in the function signature that would have been produced for the setUser method:
public function setUser(Acme\DemoBundle\Entity\User $user)
See that it requires to be of type Acme\DemoBundle\Entity\User. How do we overcome this, and have it accept any of the 4 Entities? The solution is to have all Entities be derived from a parent class. Then make the function type hint at the base class:
public function setUser(Acme\DemoBundle\Entity\Base $entity)
The base class will contain all common elements, notably a 'name', and as array collection of Reports:
// src/Acme/DemoBundle/Entity/Base.php
// ...
use Doctrine\Common\Collections\ArrayCollection;
class Base
{
// ...
/**
* @ORM\Column(name="name", type="text")
*/
protected $name
/**
* @ORM\OneToMany(targetEntity="Report", mappedBy="baseEntity")
*/
protected $reports;
public function __construct()
{
$this->reports = new ArrayCollection();
}
// ...
}
and then for each child, for example a User and a Video:
// src/Acme/DemoBundle/Entity/User.php
// ...
use AcmeDemoBundle\Entity\Base;
class User extends Base
{
/**
* @ORM\Column(name="firstname", type="text")
*/
protected $firstName;
// ...
}
and the Video
// src/Acme/DemoBundle/Entity/Video.php
// ...
use AcmeDemoBundle\Entity\Base;
class Video extends Base
{
/**
* @ORM\Column(name="title", type="text")
*/
protected $title;
// ...
and change our Report Entity:
// src/Acme/DemoBundle/Entity/Report.php
// ...
class Report
{
// ...
/**
* @ORM\ManyToOne(targetEntity="Base", inversedBy="reports")
* @ORM\JoinColumn(name="base_id", referencedColumnName="id")
*/
protected $baseEntity;
// ...
}
Remember to run the doctrine command to generate the setBaseEntity method. When you do, notice that it will now accept any class derived of Base
Then, to put on a Report on a Video for example, we get the Video, create a Report, and add the Video to the Report:
$video = // get the video you want
$report = new Report();
$report->setBaseEntity($video);
To retrieve all Reports belonging to a Comment, we get the Comment, and get the Reports:
$video = // get the video you want
$reports = $video->getReports();
foreach($reports as $report){
$reportText = $report->getText(); // assuming the Report has a `text` field
}
Update:
The inheritance relationship between these Entities can be modelled in the database with Doctrine using Single Table Inheritance:
/**
* @ORM\Entity
* @ORM\Table(name="base_entities")
* @ORM\InheritanceType("SINGLE_TYPE")
* @ORM\Discriminator(name="entity_type", type="string")
* @ORM\DiscriminatorMap({"user" = "User", "comment" = "Comment", "video" = "Video", "channel" = "Channel"})
*/
|
{
"pile_set_name": "StackExchange"
}
|
"Last orders, please." " Shaun." " Yeah?" " Do you see what I'm saying?" " Yeah, totally." "I know he's your best friend but you do live with him." "I know..." "It's not that I don't like Ed." "Ed, it's not that I don't like you." "It's all right." " It'd just be nice if we could..." " Fuck." "...spend a bit more time together." " Bollocks." " Just the two of us." " Cock it." "With Ed here, it's no wonder I always bring my flatmates out and that only exacerbates things." " What do you mean?" " You guys hardly get on, do you?" "No, what does "exacerbate" mean?" "It means to make things worse." "Right, right." "Well, I mean, it's not that I don't like David and Di, you know." " Guys, it's not that I don't like you." " It's all right." "And I do want to spend time with you." "It's just Ed doesn't have too many friends." "Can I get any of you cunts a drink?" "Anybody?" "No, thank you." "I know... friends are important to Shaun but you must set some quality time aside for yourselves." "Daffs is always taking me to see these listed buildings and I'm always dragging him to the theatre." " I'm not so hot on theatre." " How about a meal?" " It's your anniversary soon." " It was last week." " Did you do anything special?" " We came here." "Shaun." "Hog lumps." "Shaun, what I'm trying to say is I need something more, more than spending every night in the Winchester." "I want to do more interesting stuff." "I want you to want to want to do it." "Listen to me." "I'm beginning to sound like your mum." " Not that I'd know." " You haven't met his mum?" " Not yet." " Don't you get along with your mum?" " It's not her..." " Are you ashamed by your mum, Shaun?" " No, I love my mum." " I love his mum." " # She's like butter #" " Ed!" " Shaun!" " Liz." "I understand what you're trying to say and I agree." "We should get out there." "We'll start tomorrow." "I'll book a restaurant." "The place that does all the fish." "Just the two of us." "Things will change." "I promise." "Really, Shaun?" "Time, gentlemen." " # This town is coming like a ghost town #" " Shaun?" "'Player two has entered the game.'" "Haven't you got work?" "'Player two has left the game.'" "The front door was open again last night." " I'm not saying it was you." " I know, man." " I'm saying it was Ed." " Right." "I can't live like this." "Look at the state of it." "We're not students any more." " Pete..." " He brings no money into the house." " He brings a bit." " What?" "Dealing drugs?" "He sells a bit of weed now and again." "You've sold puff." "Yeah, once." "At college." "To you." "I've known him since primary school." "I like having him around, he's a laugh." "Because he can impersonate an orang-utan?" " Fuck-a-doodle-doo." " Leave him alone." "I admit, he can be pretty funny on occasion." "Like that time we stayed up all night drinking apple schnapps and playing Tekken 2." " When was that?" " Five years ago." "When's he going home?" "Hey, man, listen..." "Top left." " Uh-huh." " I was gonna say..." "Reload." " I'm on it." " Erm, since..." "Nice shot!" "Thanks." "Two seconds." " All right, Noodle?" " 'Got anything?" "'" " No, man." " 'Eball says you was holding.'" " I ain't got nothing." " 'Nothing at all?" "'" "I've only got an 'enry meself." "All right, laters." " Listen..." " Ed... since you're not working at the moment could you please clean up a bit?" "And if you play the answer machine, can you take down everyone's messages?" " Yep." " It's not that taxing, is it?" "Writing something on a scrap of paper?" " Nope." " Right." "Dom, hi..." "Come on, it was pretty funny." " Will you do what he said?" " I ain't doing nothing for him." "Do it for me, then." "I'm sorry, Shaun." " It's all right." " No, no." "I'm sorry, Shaun." "Oh!" "Oh, my God, that's rotten." " I'll stop doing 'em when you stop laughing." " I am not laughing." " Get that." " You get it." "'Shaun, it's me." "I'm gonna be a bit tied up today 'so can you book the table for eight rather than seven?" "'I'll try you at work." "Bye, bye.'" "Hey!" "You're dead." " Spare some change, please, mister?" " Yeah." "Watch out." "Are you blind?" "'The American deep space probe Omega 6 'unexpectedly re-entered Earth's atmosphere over England 'and broke apart over... '" "Nelson?" "Nelson?" "Hello, my friend." "No beer today?" "No, it's a bit early for me." "Gather round, everyone." "Gather round, please." "Gather round." "As well as Mr Sloman being off today" "I'm afraid Ash is feeling a little bit under the weather, so I will be taking charge as the..." " Oldest." " Senior staff member." "So if we can all pull together..." " Hello?" "Hello, mate." " Noel?" "I was totally munted last night." "Yeah, I spoke to him." "He's only got Henry." "All right, mate." "Laters." " Continue." " Thank you." "Erm, as Mr Sloman always says, there's no I in team but there is an I in pie." "In..." "There's an I in meat pie." "The anagram of meat is team..." "I don't know what he's talking about." "Phones off." "It's not a social gathering." " All right, keep your hair on, grandad." " Hey!" "I'm 29 for Christ's sakes." " How old are you?" "20?" "21?" " 17." "Hey, well..." "I know you don't wanna be here forever." "I got things I wanna do with my life." "When?" "You got red on you." "Hello, mate." "This one comes with a basic digital package." "You've got your lifestyle channels there, a bit of Trisha." "Er, you've got entertainment." "I don't know what that is." "News." "All the basic, er... news..." " channels." " 'Bizarre events... '" "'Coming after me... '" " Shaun." " I'm with customers." "It's your dad." "He's not my dad, he's my stepdad." "Philip." "Shaun." "You haven't forgotten about tomorrow?" " No." " Your bimonthly visit." "No, I haven't forgotten." "You could bring the flowers you forgot on Mother's Day." "I was gonna." "And not a cheap posy from a garage forecourt." "I wasn't gonna." "Well, we look forward to seeing you tomorrow." "OK." "You've got red on you." " I thought it wasn't a social gathering?" " What?" "How comes you're allowed to speak to your dad?" "A, he's not my dad, OK?" "He's my stepdad." "And B, it was an emergency." "What, like buying your mum some flowers?" "Noel, I do not find it difficult to keep my work and my social life separate." "Shaun, it's Liz for you." " Hello." " 'Hello, it's me.'" " Hello." " 'Did you get my message?" "'" " Yep." " 'So it's all OK, then?" "'" " Yep." " 'Eight at the place with all the fish?" "'Cool, that's great." "Just ring me later.'" " Goodbye, Liz." " 'Bye." "Bye, bye, bye.'" "Liz from Head Office." "It's nothing to panic about." "Sir?" "Sir?" "Sorry." "Erm, they're for my mum." "Mum. "To a wonderful Mum" or "Supermum"?" " Er... the first one." " Right." "'..." "London seems to have gone completely bonkers... '" " Oh, my God!" "Shaun!" " Yvonne." "Oooh!" "How are you doing?" " Surviving." " Are you living round here?" " Yeah, are you?" " Yeah, I just bought a place." " Bought?" " I know, bit grown-up, eh?" " Are you still with...?" " Liz." "That's great." "Glad somebody made it." "How long's that been now?" " Er, it's three years ago last week." " Did you do anything special?" " We're going out for a meal tonight." " Anywhere nice?" "Yeah..." "Come on!" " Are you gonna thank me, then?" " For what?" "Tidying up." " It doesn't look tidy." " I had a few beers when I finished." "'Hello, Fulci's." "Can you hold, please?" "'" " Do you want your messages?" " What?" "Your mum rang about tomorrow night." "Liz rang about the two of you eating out tonight." "Then your mum rang back to see if I wanted to eat her out tonight." " 'Hello, Fulci's.'" " Hi, I know it's short notice but could you possibly do me a table for two for tonight at about eight?" "'Sorry, we just gave away the last table.'" "Wasn't true about your mum." "Where are we gonna go?" " The Winchester." " Don't be stupid!" "They don't do food." " There's a Breville out back." " Ed, this is serious!" "Hello." " 'Hello, it's me.'" " Hello." "'What's the plan, then?" "'" "Erm, there's been a bit of a..." "a mix-up with the table, babe." " 'What do you mean?" "'" " They're full up." "'But..." "I thought you said it was all OK?" "'" "Yeah." "'You didn't book it, did you, Shaun?" "'" " No." " 'So... what is the plan, then?" "'" "The Winchester?" " 'Hello?" "'" " Dianne, can you let me in?" " 'I'm not sure, Shaun.'" " Oh, come on." "'Lizzie doesn't want to see you.'" " Just open the door!" " 'She doesn't want to see you.'" "Do you want me to climb up the wall, come through the window?" "Because I will." "'You're not coming in.'" " OK, see you in a minute." " 'Don't... '" "'Hi, it's me again.'" "Can we just talk about this?" "Let's go out." " To the Winchester?" " Do you want to?" " No, I don't fucking want to!" " OK." "It's not the only place in the world." "We'll go to the Shepherd's, they do Thai in there." " I'm going out with David and Di." " Let's go together." "You hang out with my friends?" ""A failed actress and a twat"?" "That's harsh." "I did not call Dianne a failed actress." "Let's just go somewhere and have a laugh." "Things'll be fine, I promise." "You promised you'd stop smoking." "You promised you'd try drinking red wine instead of beer!" " You promised we'd go on holiday." " We went to Greece!" " We met in Greece." " At a rave." "It's not the same." "You promised things would change." " You promised us free cable." " I'm working on that." "I can give up smoking whenever I want." "See, don't need them." "W" " What was the next one?" "It's not enough, Shaun." "'We appear to have a breakdown in communication there." "'Lf you or someone you know may... '" "I'd say your nine lives were up, Shaun." "Get fucked, four-eyes!" "You go out with her, you love her so much." "What do you mean by that?" "'... the devastation really is extensive." "It's a total mess.'" "I don't know what he meant by that." "Got you these." " "To a wonderful mum"?" " Oooh!" "Yeah, that's because..." "I thought it would be funny because of what you said last night about you don't want to be my mum and that." "It's just a little joke, just sort of spur of the moment." " They're for your mum, aren't they?" " Yeah." "Smooth." "If I don't do something, I'll end up in that pub for the rest of my life like those other sad old fuckers wondering what the hell happened." "What do you mean do something?" "Fuck her." "You got your pint." "You got your pig snacks." "What more do you want?" "Oh." "You want your favourite monkey?" "Shall I do Clyde?" "See, I knew you'd get over her." "# If you leave me now... #" "Who the hell put this on?" "It's on random." "For fuck's sake." "John, yes please, mate." "You know what we should do tomorrow?" "Keep drinking." "A Bloody Mary first thing, a bite at the King's Head, couple at The Little Princess, stagger back here and bang..." "back at the bar for shots." "How's that for a slice of fried gold?" " No." " Come on, man." "Talk to me." "She said if she stayed with me she'd end up in here for the rest of her life like these sad old fuckers wondering what the hell happened." "That is harsh." " These are rich, interesting characters." " Like who?" "Snakehips." "Always surrounded by women." "He's a bigamist." "Strangled his first wife with a draught excluder." "And invented the mobile disco." "What about her, then?" "Ooh!" "Cockacidal maniac." "She's an ex-porn star." "They say she starred in the world's first interracial hardcore loop." "Cafe Au Lait..." "Pour Vous." "Yeah." "What about John, then?" "He's North London Mafia." "It's true, Big Al says so." "Yeah, Big Al also says dogs can't look up." "Think about it." "Handy with a blade." "Gruff demeanour." "Bernie, the trophy wife." "He's connected." "Why is there a rifle above the bar?" " The pub's called the Winchester." " Exactly." "See, you don't need Liz to have a good time." " Don't, man." " No." "Oi, look at me." "Can I just say one thing?" "I'm not gonna say there's plenty more fish in the sea," "I'm not gonna say if you love her let her go and I'm not gonna bombard you with clichés but what I will say is this... it's not the end of the world." "Sorry, we're closed." "Pisshead." "# Ooh, white lines" "# Visions, dreams of passion" "# Going through my mind" "# And all the while I think of you..." "Oh?" "What's the matter, love, haven't you had your tea?" "# Something of a phenomenon telling my body to come along" "# Cos white lines" "# Blow away..." " Shh!" "Blow!" "# Ah, get higher, baby" "# Ah, get higher, baby, ah, get higher, girl" "# Ah... #" "What?" "He should say base." "Or freeze." "What a tit." "# You wanna dance?" "#" "Don't scratch it!" "Wait..." " That was the second album I ever bought!" " It's four in the fucking morning!" " It's Saturday!" " No, it's not." "It's fucking Sunday and I've got to go to fucking work in four fucking hours because every other fucker is fucking ill!" "Now can you see why I'm so fucking angry?" " Fuck, yeah!" " Hey!" "Pete, look," "I'm sorry, we had a couple of drinks, we split up with Liz tonight." "Just keep it down, yeah?" "Prick." "What was that?" "Nothing." " Stop it, we're friends!" " He's not my friend." "He's a fucking idiot." " What's that supposed to mean?" " It means why don't you fuck off!" "You wanna live like an animal, live in the shed, you thick fuck!" " Leave him alone!" " Stop defending him." "All he ever does is hold you back." "Or is it easier having someone around who's more of a loser than you?" " What's that supposed to mean?" " You know what I mean." "I assume it was Liz who did the dumping?" "Sort your fucking life out, mate." "What's up with your hand, man?" " I got mugged on the way home from work." " By who?" " Some crackheads." "One of them bit me." " Why?" "I don't know, I didn't stop to ask them!" "I've got a splitting headache and your stupid hip hop isn't helping." "And the front door is open." "Again!" "It's not hip hop, it's electro." "Prick." "Next time I see him, he's dead." "'Shaun, it's me." "I'm gonna be a bit tied up today 'so can you book the table for eight rather than seven?" "'I'll try you at work." "Bye, bye.'" "'Hello, pickle." "It's me, Mum." "'Dad mentioned you might be visiting tomorrow which would be lovely." "'Will you be bringing Elizabeth with you this time?" "'Only we can't wait to meet her finally 'and also I was wondering if she wanted anything special for lunch?" "'Cos these days a lot of people don't eat meat.'" "Do you want anything from the shop?" "Cornetto." "Nelson, have you got any papers?" "Nelson?" "I haven't even..." "I owe you about 15p." "Oh dear." "Sorry, I haven't got any change." "I didn't even have enough at the fucking shop." "Oh dear." "'... no official comment but religious groups are calling it Judgement Day." "There's... ' #... panic on the streets of London... #" " '... number of reports of... ' - '... serious attacks on... '" " '... people who are being... ' - '... eaten alive.'" "'Witness reports are sketchy." "One unifying detail seems to be 'that the attackers appear to be... ' '... the sensational chart-topping... ' '... or ideological connection between those committing the atrocities... '" "There's a girl in the garden." " '... emergency services... '" " What?" "In the garden, there is a girl." "'... the armed forces will be called in to provide backup and assistance." "'Scientists are still trying to establish the nature of the phenomenon 'and are unsure as to the... '" "Excuse me." "Excuse me." "Hello." " What are you doing?" " Oi!" "Oh, my God." "She's so drunk." "How much have you had, love?" " Oh!" "I think she likes you." " Shut up." " She wants a cuddle!" " I've just come out of a relationship." " Ed, do something!" " Wait there." " Ed!" " Two seconds." "I'm really flattered and everything but..." "And, hold it there." "Ed!" "Just get her off me." "Jesus!" " What's up with her eyes?" " Now, seriously..." "Mary, I'm warning you, I'll have to get physical, I mean it!" "This is it..." "Look, just fuck off!" " I think we should go back inside." " OK." " Shaun, what's going on?" " Shit, it's engaged." " How about an ambulance?" " It's engaged, Ed." " Fire engine?" " It's one number and it's busy!" " And what do you want a fire engine for?" " Anything with flashing lights." "Are they still out there?" "Yeah." "What do you think we should do?" "Have a sit-down?" "'There are reports of chaos on the motorways 'as thousands of people attempt to flee the cities." "'The M1, the M3, M4, M6... '" " Do you think this is the same thing?" " '... are all at a standstill." "'The Home Office is urging people to stay in their homes and await further instruction.'" "'Ensure all residences are secure 'with all doors and windows firmly locked and barricaded." "Police... '" " For God's sake!" "He's got an arm off!" " Get him!" "'More reports just in confirm that in all cases the attackers seem slow and shambolic... '" "I'm gonna shut the front door." "'... regarding the best way to neutralise the attackers." "'Ln the meantime, contact with these assailants is highly inadvisable.'" " Did you see his head go?" " 'Lf you're confronted... '" " We should get out there." " No, we should stay inside." " We can take 'em." " The man said to stay indoors." " Fuck the man." " In here, we're safe." "'... the attackers can be stopped by removing the head or destroying the brain." "'I'll repeat that, by removing the head or destroying the brain.'" "Hey!" "Aim for the head." " Now what?" " Get more stuff." " What's in the shed?" " Dunno, it's locked." " Why is it locked?" " It's always been locked." "No!" "That's the second album I ever bought!" "Some of these are limited..." "Ah!" "What was that?" " Blue Monday." " That was an original pressing." "For fuck's sake." " Erm..." "Purple Rain?" " No." " Sign O' The Times?" " Definitely not." " The Batman soundtrack?" " Throw it." " Dire Straits?" " Throw it." " Er..." "Ooh, Stone Roses?" " No." " Second Coming?" " I like it." "Ah!" "Sade?" " That's Liz's." " She dumped you." " I'm going to the shed." " You said it was locked." " Which one do you want?" "Girl or bloke?" " Er... the first one." "'People receiving bites have experienced headaches and nausea 'and developed symptoms... '" "You've got red on you." "'Lf you know someone who has been bitten 'it is absolutely essential that you isolate them immediately." "'The public are being strongly advised not to approach anyone... '" "Pete?" "Pete?" " Why don't we just go up?" " Don't go up there!" " Why not?" " Because A, he might be one of them and B, he might still be annoyed." "Pete?" "Maybe he went into work." "How come he didn't drive?" "His keys are still here." "Maybe he got a lift." "He said he wasn't feeling well." "Pete?" "Oi, prick!" "He's not in." " You got a fag?" " No, I've given up." " Since when?" " Since..." "Come on, babe." "Oh, she's engaged." " That was quick." " Ed, this is serious." "Liz!" " 'Shaun.'" " Mum!" "I was gonna call you." "Are you OK?" " 'Yes.'" " Are you sure?" "'Some men tried to get into the house.'" " Are they still there?" " 'I'm not sure." "We shut the curtains.'" " Did you try the police?" " 'Well, I thought about it.'" " Are you OK?" "Did they hurt you?" " 'No, I'm fine.'" " Mum?" " 'Well... they were a bit bitey.'" " Mum, have you been bitten?" " 'No, but Philip has.'" " Oh, OK." " Has she been bitten?" " No, Philip has." " Oh, OK." " Mum, what state is he in?" " 'He's fine." "Bit under the weather.'" " I see." " What's the deal?" "We may have to kill my stepdad." "Mum, you're not safe there." "I'm coming." " 'I don't want to cause a fuss.'" " We're coming to get you, Barbara!" "So, what's the plan?" "Right." "'We take Pete's car, we drive over to Mum's, 'we go in, we take care of Philip.'" "I'm so sorry, Philip." "'Then we grab Mum, go over to Liz's place, 'hole up, have a cup of tea 'and wait for all this to blow over.'" " Why have we got to go to Liz's?" " Because we do." " She dumped you." " I have to know if she's all right." " Why?" " Because I love her." "All right, gay." " I'm not staying there, though." " Why not?" "If we hole up, I want to be somewhere familiar," "I wanna know where the exits are and I wanna be allowed to smoke." "OK..." "'Take Pete's car, go round Mum's, 'go in, deal with Philip.'" "Sorry, Philip." "'Grab mum, go to Liz's, pick her up, bring her back here, 'have a cup of tea and wait for all this to blow over.'" "Perfik." "No, wait, we can't bring them back here." " Why not?" " It's not exactly safe, is it?" "Yeah." "Look at the state of it." "Where's safe?" "Where's familiar?" "Where can I smoke?" "'Take car, go to Mum's, kill Phil.'" " Sorry." " 'Grab Liz, go to the Winchester, 'have a nice cold pint and wait for all this to blow over.'" "How's that for a slice of fried gold?" "Yeah, boy!" "'To recap, it is vital that you stay in your homes." "'Make no attempt to reach loved ones 'and avoid all physical contact with the assailants.'" "Do you believe everything you hear on TV?" "Yeah, I can see it, it's just outside." "Any zombies out there?" " Don't say that." " What?" " That." " What?" "That." "The Z word." "Don't say it." " Why not?" " Because it's ridiculous." "All right." "Are there any out there, though?" "Can't see any." "Maybe it's not as bad as all that." "Oh!" "No, there they are." " OK, let's do this, yeah?" " Yeah?" "Yeah!" "I got to do a wee first." "Ooh... can I drive?" "What?" "I always wanted to drive Pete's car." "I might not get another chance." "OK." "Ahh!" "Oh, God." "Sorry, Pete." "Listen, we're gonna borrow your car, OK?" "Erm, we'll bring it back." "And if you're feeling better later, we're... we're going to the pub." "You're more than welcome to, er... join us." "Pete'd be so pissed off if he knew I was driving his car." "I don't think so." "Shaun, which button's central locking?" "Oh, cock it!" "Get lost!" "Drive." "'The Church of England has joined other extremist religious groups 'in proclaiming this "a sign of a coming apocalypse"." "'Downing Street is refusing to be drawn into a religious debate." "'The bodies of the recently deceased are returning to life and attacking the living.'" "Let's put something else on." "What are you doing?" "Ed, look out!" "I think we hit something." "Or someone." "Ow!" "Are you all right?" "Come on, let's just go." "Hello?" " He's gonna be dead either way." " That's not the point." "Oh, thank God for that." "Ah, hello." "Who's a pretty boy then?" "You didn't say Barbara had a Jag." "I've always wanted to drive a Jag." "It's Philip's." "He won't let anybody near it." "I put a Mars Bar in it once and he chased me with a bit of wood." " Fuck, it's gorgeous." " It seems pretty clear, we should go." "Why don't I stay here?" "What, what, what, what, what?" "You know, just look after the car." "If there's any problems..." "I'll honk three times." "OK." "Don't forget to kill Philip!" " Hello, pickle." " Hi, Mum." "You've got red on you." "Yeah." "Mum, is everything OK?" "Are you all right?" " Yes, I'm fine." " Where's Philip?" " Dad's in the lounge." " He's not my dad." "Really!" "Ed's outside, we'll take you somewhere safe." " What about the doctor?" " I don't think he's gonna come." " Philip won't want to leave the house." " Philip isn't..." "Put the kettle on and I'll deal with Philip." " OK." " OK, then." " You hungry?" " Not really." " I'll make some sandwiches." " OK." "No fighting, you two." "'... you may ask where is our God?" "'But throughout the land, men and women of integrity 'rise up to confront the unseen enemy." "'They will not be cowards, 'they will be courageous in the face of danger.'" " I'm so sorry, Philip." " Why?" " What have you done now?" " Nothing." "What have you got there?" "Flowers, I hope." "No, it's... cricket bat." " Is that for the jumble, Shaun?" " No." "Yes." "Your toys are upstairs." "I was going to give them to the jumble." " I took them all to the tip, Barbara." " What?" " Shaun wants to take us somewhere." " Don't be silly." "Maybe you should stay here and wait for the doctor." " You said he wasn't coming." " You didn't call the doctor?" " Just to be on the safe side." " I'm fine." "I ran it under a cold tap." "We had ourjabs when we went to the Isle of Wight." " But, Philip..." " It's overblown nonsense." " A lot of drug nuts running wild." " But I should get Mum away." " I'm not going anywhere without Philip." " Mum, look!" "How about that tea?" "Mum, how much do you love Philip?" " Two sugars, is it?" " I haven't had sugar in my tea since 1982." "Oh, yes." "Will you cut me some bread, love?" "Mum, look, how much do you love Philip?" "Do we have to go through all that again?" "What would you say if I told you that Philip's been quite unkind to me?" "You weren't easy to live with." " He chased me with a bit of wood!" " You did call him a you-know-what." " Did he tell you that?" "Motherfucker." " Shaun!" "Sorry, Mother..." "Mum!" "Did you know that on several occasions..." "he touched me?" "That wasn't true." "Made it up, shouldn't have, sorry." " You don't understand..." " No, you don't understand." "Philip is my husband and has been for 17 years and I would at least expect you to respect my feelings." "You must be more adult about these things." "Yeah." "Come on, Shaun." "There comes a time when... you just... gotta be a man." "I'm telling you, it's a fuss over nothing." "Hello, Edward." " Hi, Barbara." " My, how you've grown." "You better believe it." "What happened?" "Why is he coming?" " It wasn't as easy as all that." " Are we following you?" "No, it's probably best if we..." "Where's the car?" "Hmm..." "I pranged it." "You were parked." "Yeah." "I guess we'll have to take the Jag." " Philip?" " What?" "Look out!" "Christ!" " Give me the car keys!" " You're not driving that car." "Give me the car keys!" "Ed, get Mum in the car." "I'm on it." "Don't worry about me." "Worry about your mother." " Maybe I should drive." " I've adjusted the seat now." "Just be careful!" "Can you please turn that noise down?" "Are you all right, dear?" "Would you like another tissue?" "I'd be fine if it weren't for that bloody racket." " It is a bit loud." " Sorry, Barbara." "You do realise this is a 20 mile an hour zone?" "Oh, yeah." " Here." " Mum, this won't take a second, all right?" " Have you still got the child locks on?" " Safety first, Shaun." "You can't..." "Hey, man, watch the leather." "Keep moving." "If there's any problems..." "I'll do the honks." "'Hello?" "'" "Hello?" " Guys, can you let me in?" " What are you doing?" "I've come to get you." " What are you doing?" " I've come to take you somewhere safe." " We were safe!" " You don't know that." "There was two of them out there." "How many are there now?" " Lots." " Lots!" "It only takes one of them to know you're in here." "I tried calling but I couldn't get through." "I had to come up with a plan." " Oh, you made a plan?" " We have to get out of here." "If we don't, they'll tear us to pieces." "And that will exacerbate things for all of us." " Shaun..." " This is not about you and me." "This is about survival." "We need to be somewhere more secure, somewhere we can stay alive." " We're with you, Shaun." " What?" " We're with you." " I'm not going out there." "Daffs!" "I don't want to be torn to pieces and if you think about it, neither do you." " I'm not going if Lizzie's not going." " She's going." " I'm not staying by myself." " You're going, aren't you?" " Lizzie?" " Liz?" " All right, Flash, what's your plan?" " Great." "I've got a car but it's gonna be a bit cramped now." "Has anyone got transport?" " Yes!" " Great." "Where?" "Oh, no." "Well, I've passed my test." "David?" "I don't see the point of owning a car in London." "Look, just get any blunt objects together." "If you get cornered, bash them in the head." "Keep together, stay sharp and follow me." "So, what's the plan, then?" "We're going to the Winchester." "OK?" "Let's go." " Come on." " Where's the car?" "What's up, niggaz?" "Apologies for the cramped conditions." "Di, stay on David's lap." "Philip, hang in there, OK?" "Ed, get there fast." "And, Mum, this is Liz." "Liz, Mum." "Hello." "Ed, be careful." " Is your dad all right?" " He's not my dad." " He's bleeding." " I know!" "Ed, be careful." "Thought you wanted to get there fast?" "I want to get there in one piece!" "Why are we going this way?" "Chill out, it's a short cut." " That road goes to the Winchester." " Shaun..." " Liz, please!" " I don't understand why the Winchester?" " It's a pub, it's safe, it's secure." " They know us." " What makes it so secure?" " It's got big heavy doors and deadbolts." "And there's a rifle above the bar." " I would think that's deactivated." " It's not." "John's connected." "Big Al says so." "Big Al also says dogs can't look up." " They can't." " Can't they?" " Of course they can!" " Are you sure?" "Yes!" "The pub is the right place to go." "Everything is gonna be fine." " Shaun..." " I turned it off, all right?" "What?" " Being a father, it's not easy." " What?" "You were 12 when I met you, you'd already grown up so much." "I just wanted you to be strong and not give up because you lost your dad." " Philip, you don't have to explain." " No, I do." "I always loved you, Shaun." "And I always thought you had it in you to do well." "You just need... m-motivation." "Somebody to look up to and..." "I thought it could be me." "Would you just... take care of your mum?" "There's a good boy." " Ed, can you pull over?" " Two seconds." " Ah!" "Ed, please pull over." " What for?" " Just pull over!" " All right." "Whoa, momma!" " What the hell are you doing?" " Chill out." "Everyone's all right." "Stop telling me to chill out!" " And everybody is not all right." " What's wrong?" "Mum, it's Philip, he's gone." " Where's he gone?" " Mum, he's dead." "No, he isn't." "Get out!" "Jesus!" " Get it open!" " Ed, get the child locks!" "Quick!" " The child lock!" " Ed!" " Now what?" " We keep moving." " How do we get Philip out the car?" " We haven't got time." " We can't leave your dad." " He's not my dad!" "Mum, he was, but he's not any more." " I really think we could..." " That's not your husband in there." "OK?" "It looks like him but there is nothing of the man you loved in that car." "Let's go, shall we?" "You're not suggesting we walk?" " Everybody OK?" " Yeah, we're having a whale of a time." "Maybe we should all hold hands, make a crocodile?" "Do you know where we're going?" " David, come on." " Sorry." " Do you, though?" " Yeah, totally." "Are we close?" "We're pretty close." " Is your mum gonna be OK?" " Yeah, I hope so." " Did you contact your folks?" " They're in Antigua." " Do you think this is happening there?" " I hope not." "About last night." "I know loads of things were said but..." " You said this wasn't about us." " No, it's not." " You said it was about survival." " It is, but..." "You're notjust trying to prove something?" "I meant everything I said last night." " Maybe this isn't the best time to talk." " OK." "Shaun." "Yvonne." " How's it going?" " Er, surviving." "Hey, Liz." "Long time no see." "Hey, Yvonne." "Erm... this is my boyfriend, Declan." "Erm... my friends Mark, Maggie," " my mum and my cousin Tom." " All right." "This is Shaun and his girlfriend Liz." " They've split up, actually." " Oh, really?" " Yeah." " That's a shame." "We should scoot." "Have you got somewhere you're going?" " Yeah, we're going to the Winchester." " The pub?" " Yeah." " Right, well." "Good luck!" "Hi." "Hello." "Hi." "Hello." "All right." "See?" "The Winchester's just over there." " Over where?" " Over there." "Just over there, over the 20 garden fences?" "What's the matter, David?" "Never taken a short cut before?" "Shaun." "I'm sure Jill and Derek live round here." "Now, the Winchester is just over that fence, all right?" "Everyone stay where they are, I'll..." "Are we all here?" " One, two, three, four, five." "Yes." " There's six of us." "Mum!" "I'm coming!" "Mum, are you OK?" "Are you OK?" "I don't think Jill and Derek do live here." "Shaun?" "Are you all right?" "What are you doing?" "Shaun!" "Yes!" "No, that's rubbish." " Ooh!" " Yes!" "Feel free to step in any time." " You did all right." " I didn't want to cramp your style." "I'm so sorry about that." "Right." "OK, everyone stay where they are." "I'm gonna check if the coast is clear." " Is it clear?" " No." " How many?" " Lots." "Oh, well, that is just great." "There are a fair few of them, yes." "I trust Shaun has another genius scheme up his sleeve?" "This is hardly constructive, David!" " We should have stayed at the flat." " Why didn't you?" "Because..." "Because, because of..." "Captain Wow." " Will you stop it?" " We'll get nowhere moaning." "Right, let's all shake out." "Get nice and limber." "Or not." "Take another look at the way he moves." "Remember, very limp." "Almost like sleepwalking." "Look at the face, it's vacant with a hint of sadness." "Like a drunk who's lost a bet." "OK, let's try, shall we?" "Liz." "Nice." "Good vocal work." "OK." "Barbara, that's excellent." "Sorry, dear, I was miles away." " D-Daffs." " Uh." " Come on!" " Urgh." "It's mournful, sorrowful, you're dead and you hate it." "Go." "Uuurgh." " Much better." "Ed." " I'll do it on the night." "This is the night!" " What is that?" " What about yours?" "Who died and made you fucking king of the zombies?" "Don't..." "Yeah, that's pretty good." "Right, let's all try together, shall we?" "One, two, three." "Go, go, go, go." "Shit!" "Get round me, get round me!" "John?" "John, Bernie, are you there?" "It's Shaun and Ed." "We come in all the time." " They can hear you." " We do the quiz." " They know." " Stay in character!" " I can't see them." " Shaun, we have to get inside." " Break the window." " We'll be exposed!" " Break the window!" " Is there another way in?" "Yeah, there's..." "Two seconds." "Hey, Noodle." " 'All right?" "'" " Yeah, man, I can't talk now." " 'You get anything yet?" "'" " I got nothing." "I'm in the middle of something." "Yeah, it's weird." " Oi!" "What are you doing?" " What are you doing?" "You stupid moron." " Fuck off." " You fuck off!" "Fuck fucking off!" "I've spent an entire life..." "Look at me." "I've spent my entire life sticking up for you and all you ever do is fuck things up and make me look stupid." "I'm not gonna let you do it any more." "Not today!" "Shaun." "Oh." "You were saying about another way in?" "Yeah, there's a..." "What are you doing?" " No!" " Get inside." "No, they'll follow us." "Keep up the act, I'll have to do something." " What do you mean do something?" " Hello, hello, over here!" "Come on, that's it." "Come on, that's it, over here!" "Over here!" "Come on." "Oh, bollocks." " Wait till I've gone then get inside." " Gone where?" "I won't be long, I promise." "Come on, you pasty-faced fuckers, let's go!" "Follow me." "Come on, this way, that's it!" "Come on." "This way, follow me." "Come on, follow me." "Come and get it." "It's a running buffet." "All you can eat!" "Shall we go in, then?" " Come on, Barbara." "Leave those." " I think they're for me." " He's been gone a good 20 minutes." " 17." "Either way, we need to think about blocking up this window." "We're totally exposed." " How will Shaun get back in?" " I'm sure he'll knock." " Let's get some lights on." " David, wait!" " Power's off." " That's a good thing." " We mustn't advertise our whereabouts." " No, of course not." " Do we, David?" " No, of course not, Lizzie." "I'm just being proactive." "There's no lights, no power." "And there's a hole in the window." " You did that, you twat." " Somebody had to do something!" "We were in a spot of bother." "Somebody has to take control of the situation and if none of you are prepared to accept that responsibility then perhaps..." "I should." " Will Shaun be gone long?" " He'll be back soon." " How do you know?" " I don't think he'd leave us." "Wouldn't he?" "How can you trust a man you binned for being unreliable?" "A man whose idea of a romantic nightspot is also an impenetrable fortress." "This is a pub!" "We are in a pub!" "What are we going to do?" "I could get a round in." "Let's just wait for Shaun." "We'll barricade the window when he gets back." "What then?" "How long?" "Days?" "Weeks?" "A month?" " What are we going to eat?" " Toasties!" " There's a Breville out back." " Great." "Saved by nibbles." "Must be why Shaun took us here before he left." " He's coming back!" " Why?" "Because he promised?" "If he does, do you think his master plan will be anything more than eating peanuts?" "Is he just gonna stroll in and suddenly everything's OK?" "I don't know, David!" "I don't know any more than you do." "But what I do know is that we're here now and we have to make the best of it." " Ed, get me a double vodka." " Right you are." "I'll have a drink too." "Would you like a drink, Barbara?" " Hello." " Right, great, fuck it." "We'll have a party and get completely smashed." "We've got our nibbles." "We've got our Mini Cheddars." " David!" " We got Twiglets." "Oh, look, hog lumps." "Pickle!" "Hello, Mum." "You all right?" " Everybody OK?" " Yeah." " Any sign of John and Bernie?" " No." " Check upstairs?" " It's locked." " Phones?" " Dead." "Same as the power." " OK..." " Nice of you to join us." "Yeah, well, I promised, didn't I?" "How did you lose them?" "I just gave them the slip." "Wasn't difficult." "They're not all that." " Well, how did you get in?" " The back door." "I tried to tell you before you smashed the window." "But it wasn't me blowing our cover by arguing with my boyfriend." " He's not my boyfriend." " Might be a bit warm, the cooler's off." "Thanks, babe." "So, what's the plan, then?" "Would anyone like a peanut?" "Why don't you do your Clyde?" "Has anyone seen Every Which Way But Loose?" "Ed does the most amazing impression of the orang-utan." " No, it's not." " It is." "Come on." "Do it." " No, no." " But it's brilliant." "Come on." " You've built it up too much." " No, just do it." "Everyone wants to see it." "We could do with a laugh." "Do it, do it, do it, do it." "Come on, do it." "I'm not a performing monkey." " The power's on." " It's not, I tried it earlier." "The streetlights are on." "I'll flip the fuses, we can watch television." " What about the lights?" " Yes." "Don't advertise our whereabouts." "The lights are on a separate circuit." "Just flip the mains breakers." "It's quite simple, David." " Is that it?" " That's it." "Liz, erm... have you got a quid?" "Erm..." "Yeah." "Never anything on, is there?" "What?" " I think we might have a bit of a problem." " What do you mean?" " They followed me." " I thought you said you gave them the slip." "Yeah." "Well, we'll just have to be extra quiet then, won't we?" "What?" "Ed, no!" "Turn it off!" "Shit!" " Where the hell did he come from?" " I dunno." " Who put this on?" " It's on random." "Dianne, get Mum somewhere safe." "David, kill the Queen." " What?" "!" " The jukebox!" " Liz, Ed, grab a weapon." " What about the rifle?" " Cocktails?" " What do you mean?" "The flaming spirits Get a rag in, light it." " But the whole place could go up." " What, then?" "How about pool?" "# Don't stop me now" "# Don't stop me" " # Cos I'm having... #" " John, it's time at the bar." "# I'm a shooting star leaping through the sky like a tiger" "# Defying the laws of gravity" "# I'm a racing car passing by like Lady Godiva" "# I'm gonna go, go, go" "# There's no stopping me" "# I'm burning through the sky..." " I can't find the switches!" " Then get the fuses!" "# That's why they call me Mister Fahrenheit" "# I'm travelling at the speed of light" "# I wanna make a supersonic man out of you" "# Don't stop me now" "# I'm having such a good time" "# I'm having a ball" "# Don't stop me now" "# If you wanna have a good time" "# Just give me a call" "# Don't stop me now cos I'm having a good time" "# Don't stop me now" "# Yes I'm having a good time I don't wanna stop at all" " # I'm a rocket ship on my way to Mars..." " Why is Queen still on?" " We have a situation!" " I know!" " Fuck this." "Two seconds." " Ed!" "Oi!" "# Oh, oh, oh, oh, oh, explode" "# I'm burning through the sky, yeah..." "Shaun!" "#... call me Mister Fahrenheit..." "No!" "Yes!" "Yes!" "In the head!" "Ooow!" "Shauny!" "# Don't stop me Ooh, ooh, ooh" "# Don't stop me, don't stop me good time, good time" "# Don't stop me... #" "Why didn't you just shoot him, man?" "Ed, for the last time, it's..." "I fucking knew it!" "What did I tell you?" "Big Al was right." "OK." "But dogs can look up." " Has anyone fired a gun before?" " Oh, yeah." " Apart from Ed?" " I'm a pacifist." " Apart from David and Ed?" " What was Ed's experience?" " He shot his sister in the leg." " I think you should do it." " We should all do it." " How can five people operate one gun?" " No one has any experience." "We've got..." " 29." " 29 bullets." " Call them shells." "29 shells." "I need someone to help me reload, everyone else to look out." "I will fire." "As Bertrand Russell said the only thing that will redeem mankind is co-operation." "I think we can all appreciate that now." " Was that on a beer mat?" " Yeah, it was Guinness Extra Cold." " I won't say anything." " Thanks." " They want to come in." " Get Mum out of the way." "David, Dianne, let's go." "Ed, give me some shells." " Barbara." " Oh, hello." " Would you like to come with me?" " That'd be lovely." "It's nice to meet you, finally." "Shaun's always talking about you." " Really?" " Yeah." " Ah!" " Barbara?" "Are you all right?" "I've got something for you." "Shaun's father gave me this." "Shaun's real father, I mean." " I'd like you to have it." " Barbara, Shaun and I have..." "It seems only right." "I didn't want to say anything." "I thought Shaun would be worried." "Liz!" "Here they come!" " Stay behind me!" " There's one!" "Sorry, we're closed!" " The safety!" " That happened to me on stage." "Cock it!" " There!" " Where?" "Three o'clock!" " Oh!" "Over there again." "Quarter to twelve." " What?" " 11:45!" " Keep it simple!" "Top left!" " Reload." " I'm on it!" "Come on, come on." "Nice shot." "OK, David, Dianne, let's block up this window!" "Ed, take over a sec." " What is it?" "What's wrong?" " Shaun, I'm so sorry." " Mum, what's wrong?" " I didn't want to be a bother." " Mum, what's wrong?" " I didn't want to be a bother." " Mum, I don't understand." " The man in his pyjamas." "I didn't want to say anything, I thought you'd be upset." "No, you should've said." " Take over a sec." " What's going on?" " What's up?" " What the hell's going on?" " Barbara's hurt!" " Just take over a minute." "Jesus." "I know first aid." "Just hang on." "Mum!" " Mum." " Is she...?" "Just hold on, Mum, you're gonna be fine." " I never thanked you." " For what?" "For these. "To a wonderful mum."" "It's been a funny sort of day, hasn't it?" "Oh, no, no!" "Mum, come on." "Stop it, stop it, stop it." "Mum." "Please don't go." " What are you doing?" " We have to deal with her." " Daffs!" " For Christ's sake!" " She's gonna come back!" " She's not going anywhere." " She'll change." " She's my mum!" " She's a zombie." " Don't say that." "Move aside." "I'm gonna count to three." "One... two... three!" " Don't point that gun at my mum!" " Calm down." "Don't point that gun at Barbara!" " Ed, don't exacerbate things!" " What's that mean?" " This isn't exactly fair." " Here." " Thank you." " It's OK." "Please, can we just stop and think about this?" "Tell him to put the gun down!" "She's gonna change." "You know I'm right." "And when she does, she'll kill all of us." "That's what your ex can't seem to realise." "That's what this is about, isn't it?" "He doesn't like me!" "He's always hated me and now he wants to shoot my mum." " She's not..." " You've never thought I was good enough." " What?" " We all know you're in love with Liz." " That is not true." " Yes it is!" " That is not true." " Yes it is." " What?" " I know you only hung out with me to get close to Liz and when she knocked you back, I was there to pick up the pieces." "I've come to terms with that, Daffs, why can't you?" "Lizzie, I want you to know that my feelings for you..." "David!" "There are slightly more pressing matters at hand here." " Like "Daffs" pointing a gun at my mum." " I'm not being unreasonable, "Pickle"." " Christ, she's not dead!" " She's finished." "For a hero, you're quite a hypocrite." "You've gone from a chartered accountant to Charlton Heston." "I'm not a chartered accountant." " Well, you look like one." " Yeah." " I'm a lecturer." " You're a twat." "She's not your mum." "In a minute she'll be just another zombie." " Don't say that." " We're not using the Z word." "Please can we just calm the fuck down!" "Now, I can see what David is trying to say..." " Thank you." " Even if he is being a twat." "Yep." "And, Shaun..." "Shaun, look at me." "I can't begin to imagine how hard this must be for you but we can't afford to fall apart." "Shaun, babe..." "I know you don't wanna hear this..." "You do it." "I'm sorry, Mum." "Do it!" "Well..." "I think we're all agreed, you did the right thing there." "David, no!" "David!" " Right, I'm leaving." " What?" " I'm not staying here." " David, it's suicide." " I think you should go." " We will." "We can defend ourselves." " What do you mean "we"?" " What do you mean?" "Opening that door now would be a very silly thing to do." " You gonna stay here with him?" " Look..." " After where he led us?" " Let me finish!" "Shaun didn't ask you to come here." "You came for the same reason I did." " Get away from that door this instant!" " But..." "David!" "I'm so sorry, Dianne." "Maybe I'm not the one you should be apologising to." "Shaun..." "David!" "No!" "No!" "No!" "No!" "David!" " No!" "No!" " David!" "David!" "Dianne, no!" "I'm coming, David!" "David!" "Oh dear." " Cocktails?" " Do it." "Stay back!" "Shaun, hold them!" "Heads!" " Ed!" " Hurry up!" "Shauny, look who it is." "Fuck-a-doodle-doo!" "Ed!" " Shaun!" " Ed!" "Don't groan at me, you thick fuck!" "I can't hold them!" "No!" " Shaun!" " Pete!" "I said leave him alone!" "Ed." "Ed!" "Shaun, bar!" "Ed!" "Ed!" "Ed!" "What?" "Chuck us a cloth, will you?" " Where are the shells?" " They're on the bar." "Oh, fuck." "Ow!" "Fuck it!" "What are we gonna do?" "Where are we gonna go?" " The cellar." " Maybe we can get out!" " There's a hatch out on to the street." " You mean go back outside?" "Oh, give us a fucking break!" "Go!" "I can see the street." "Come on." "Come on!" "Why won't it open?" "That's it then, we're trapped." ""Hey, let's go to the Winchester!"" "Whose fucking idea was that?" "Man, I've really ballsed this up." " No, you haven't." " I have, Liz." "I couldn't save us." "I couldn't save Di or David." "I couldn't even save Mum." "I'm useless." "You shouldn't feel so responsible, you tried." "You did something, that's what counts." "Yeah, I suppose." "Do you think they'll get through?" "Yeah." "How many shells have we got left?" "Erm... two." "I suppose we could take a few of them out if they stand in a line." "I wasn't thinking about them." "I know." " What about Ed?" "There's only two shells." " I don't mind being eaten." "How are we gonna do this?" "I don't know." "Erm..." "One of us has to go first." "One should do the other and then do themselves." "You should do me." "I'll muck it up if I have to do myself." "I don't think I've got it in me to shoot my flatmate, my mum and my girlfriend all in the same evening." "What makes you think I've taken you back?" "Well... you don't want to die single, do you?" "Actually..." "I would like to be shot." "Besides, I've changed." "I haven't had a fag since yesterday, I promise." "He hasn't." "Maybe we should have one now." " You left them in my flat." " Yeah, in the bin." " I was desperate." " Sneaky monkey." " Sorry." " I won't say anything." "Come on, we're getting out of here." " I mightjust stay here." " But we can get out." " You two should make a go of it." " What are you talking about?" " I only hold you back." " Yeah, but..." "Maybe you should have this." " Don't you want it?" " No." "I can't hit anything with it anyway, I'm rubbish." " Ed, I'm sorry." " What for?" " Cos I was shouting at you earlier on." " It's all right." "I'm sorry too." "It's OK." "No..." "I'm sorry, Shaun." "Oh, God!" "That's not funny." " I'll stop doing 'em when you stop laughing." " I'm not laughing." "You'd better be off." "Bye, Ed." "Love you." "Cheers." "I love you too, man." "Gay." " Ready?" " Ready." "Section One." "Ready." "Zombies to your front." "Fire!" "Fire!" "Oh, my God!" "Shaun!" "Yvonne?" " How are you doing?" " Surviving." "They're taking us somewhere safe." "I thought you two might want to tag along." "Is itjust the two of you?" "Well, glad somebody made it." "Come on." "'Next, the story of ten-year-old Enrique Ramírez 'who fought off the reanimated corpses of his entire family." "'Lt's all coming right up on Zombies From Hell.'" "'Six months on and, to many, the events of Z-Day must seem like a bad dream." "'As we now know, the phenomenon resulted from the use of... ' '... not something I ever expected as a newscaster to have to say on air." ""'Removing the head or destroying the brain." Incredible.' '... to have a top band with us, talking about their work for the charity Zombaid... '" "'The fact that the mobile deceased retained their primal instincts 'make them ideal recruitment for the service industry, 'as well as other roles such as... '" "'He can't quite get it." "Ooh!" "He's fallen, he's down!" "'" "'I don't see nothing wrong with it but I know that some people would." "'But he's my husband, you know." "I still love him and I got the ring." " 'You go to bed with it?" "' - 'Course I do.'" "'People can't understand why she's with him?" "'" "'Lnitial claims that the virus was caused by infected monkeys have been dismissed... '" "Hello." "What's the plan, then?" "Right, a cup of tea... then we get the Sundays." "Head down the Phoenix for a roast, veg out in the pub for a bit then wander home, watch a bit of telly, go to bed." " Perfik." " I'll get the kettle on." " Can I have two sugars, please?" " Ooh!" "Adventurous." " I might pop into the garden for a bit." " Go on, then." "Two seconds." "'Player two has entered the game.'" "Ed!" "Yes!" "# Ooh, you make me live" "# Whatever this world can give to me" "# It's you, you're all I see" "# Ooh, you make me live now, honey" "# Ooh, you make me live" "# Oh, you're the best friend that I ever had" "# I've been with you such a long time" "# You're my sunshine" "# And I want you to know that my feelings are true" "# I really love you" "# Oh, you're my best friend" "# Ooh, you make me live" "# Ooh, I've been wandering round" "# But I still come back to you # Come back to you" "# In rain or shine you've stood by me, girl" "# I'm happy at home # Happy at home" "# You're my best friend" "# Ooh, you make me live # Ooh" "# Ooh, you're my best friend #" "# I was so tired of being upset" "# Always wanted something I never could get" "# Life's an illusion, love is a dream" "# But I don't know what it is" "# Everybody's happy nowadays" "# Everybody's happy nowadays" "# I was so tired of being upset" "# Always wanted something I never could get" "# Life's an illusion, love is a dream" "# But I don't know what it is cos" "# Everybody's happy nowadays" "# Everybody's happy nowadays" "# Life's an illusion, love is a dream" "# But I don't know what it is" "# Everybody's saying things to me" "# But I know it's OK, OK" "# Everybody's happy nowadays" "# Everybody's happy nowadays" "# Life's an illusion, love is a dream life's an illusion, love is a dream" "# Life's an illusion, love is a dream life's the illusion, love is a dream" "# Life's an illusion, love is a dream life's an illusion, love is a dream" "# Life's an illusion, love is a dream life's an illusion, love is a dream" "# Everybody's happy nowadays" "# Everybody's happy nowadays" "# Days" "# Bet you are tired of being upset" "# Always wanted something you never could get" "# Life's no illusion, love's not a dream" "# Now I know just what it is" "# Everybody's happy nowadays" "# Everybody's happy nowadays" "# Days #"
|
{
"pile_set_name": "OpenSubtitles"
}
|
Videogame Mechanics in Games for Health.
Game mechanics have been identified as "methods invoked by agents for interacting with the game world."(1) They are elements of game play that provide a primary source of interactivity and structure how videogames proceed. Many kinds of game mechanics have been generated. Some provide fun or enjoyment, others may provide excitement or even suspense (i.e., emotional aspects of game play), whereas some are guided by principles of behavior change theory. Game mechanics that succeed in off-the-shelf entertainment videogames, however, may not work in serious games, such as games for health (G4H). Although game mechanics are key to a serious videogame's success, maintaining a balance between the serious intent of the game while keeping it fun, there has been little attention to game mechanics in the academic G4H literature. Several eminent games for health developers (academics and nonacademics) were asked to share their experiences in regard to game mechanics in the serious videogames they have developed.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Q:
How to convert shape file x, y coordinate to lat long in php
Hello I am working with shape file in my yii2 project. I have created function to upload and read shape files. (.shp, .shx, .dbf).
I am getting data like :
In above image I am getting x and y cordinate. From it how can i get lat and long ?
I need to save this in database and display in map.
How to convert x and y to lat and long in php.
A:
The x coordinates are the latitudes and y coordinates are the longitudes. There is no need to convert these unless you want to work with a coordinate system that doesn't use decimal degrees.
|
{
"pile_set_name": "StackExchange"
}
|
All relevant data are within the paper and its Supporting Information files.
Introduction {#sec001}
============
*Staphylococcus aureus* (*S*. *aureus*), one of the most common Gram-positive bacterium, causes a series of life-threatening diseases including sepsis, bacteremia and necrotizing pneumonia etc \[[@pone.0136888.ref001], [@pone.0136888.ref002]\]. In recent years, the treatment has become more and more difficult for infections caused by methicillin-resistant *S*. *aureus* (MRSA), strains that are resistant to almost all the commonly used antibiotics \[[@pone.0136888.ref003]\]. This fact urges pharmaceutical companies to develop new and effective prophylactic vaccines and immunotherapies in the battle against the bacterium. However, attempts at making *S*. *aureus* vaccines have so far been unsuccessful \[[@pone.0136888.ref004], [@pone.0136888.ref005]\]. This failure has been attributed partly to the feature of the bacterium, which exhibits diverse arrays of virulence factors among individual clinical isolates, and can change its surface antigen repertoire potentially in different infection states \[[@pone.0136888.ref005]\]. Thus, it would be wise to select a conserved antigen that is expressed on bacterium persistently and steadily as a vaccine candidate. The second obstacle for development of *S*. *aureus* vaccine is that the relationship between the bacterium and the immune system remains ambiguous. *S*. *aureus* can evade the host immune system by inhibiting phagocytic uptake and its killing activity \[[@pone.0136888.ref006], [@pone.0136888.ref007]\]. Moreover, the effect of antibody-mediated opsonization is controversial \[[@pone.0136888.ref004], [@pone.0136888.ref005], [@pone.0136888.ref008]\]. Recently, more and more reports indicate that T cell-mediated immune responses, especially stimulating and secreting cytokines such as IL-17 or IFN-γ, might play an important role in prevention of *S*. *aureus* infection \[[@pone.0136888.ref009]--[@pone.0136888.ref017]\].
Peptidoglycan (PGN), accounting for about 50 percent weight of the bacterium wall, is made of polymeric meshwork of glycan strands crosslinked by short peptides \[[@pone.0136888.ref018]\]. In bacterium, PGN maintains the cell shape and osmotic pressure of cytoplasm. Therefore, it is a crucial and conservative component of *S*. *aureus*. However, as a thymus-independent antigen (TI antigen) with weak immunogenicity, PGN is not considered as a rational vaccine candidate. Mimetic peptides, acting as a thymus-dependent antigen (TD antigen) with relatively stronger immunogenicity, have been used as a promising surrogate for carbohydrate to generate vaccines. Immunization with a peptide, mimicking a carbohydrate antigen of tumor cells or polysaccharide in bacterium, could augment tumor-specific cellular responses or induce protective antibody responses to bacterium \[[@pone.0136888.ref019], [@pone.0136888.ref020]\]. Recently, we successfully obtained a series of PGN-mimicking peptides from phage peptide library by using an anti-PGN mAb as the target \[[@pone.0136888.ref021]\]. In this study, we synthesized a multiple antigenic peptide (MAP), named MAP27, which carried four copies of an identical peptide derived from the screening of the phage peptide library. Immunization with MAP27 decreased the bacterial burden in organs of the mice and significantly prolonged their survival time after *S*. *aureus* lethal-challenge. This protection is mainly correlated with the induction of IFN-γ- and IL-17A- producing T cells by MAP27 immunization.
Materials and Methods {#sec002}
=====================
Mice and Ethics Statement {#sec003}
-------------------------
BALB/c mice (4--6 weeks of age, female) were purchased from the Experimental Animal Center, Southern Medical University, Guangzhou, China. All of the animal experiments were approved by the Animal Ethical and Experimental Committee of the Southern Medical University (permit number: 2014--028). All mice were maintained under specific-pathogen-free conditions in a clean room at the Institute for Animal Experimentation in Southern Medical University. They were kept on a 12:12 h light-dark cycle with food and water provided *ad libitum*. Animal experiments were carried out in strict accordance with national guidelines for animal welfare. To alleviate pains, all mice were anesthetized with sodium pentobarbital (50 mg/kg) via abdominal cavity before *S*. *aureus* infection. According with the principle of animal ethics, the humane euthanasia was performed. During the bacterial lethal challenge, mice were observed every 2 hours for the first 48 hours, and the survival was monitored for at least 7 days. Mice that became moribund (such as hunched back, ruffled fur and lethargy), the survivors at the endpoint of observation during *S*. *aureus* lethal challenge, and the mice infected with *S*. *aureus* for three days were sacrificed by CO~2~ asphyxiation.
Peptide Synthesis {#sec004}
-----------------
All peptides, including three linear peptides (SP27, SP27' and L2), a multiple antigenic peptide (named MAP27), and the attaching backbone of MAP (MAPctrl) were synthesized by Hybio Pharmaceutical (Shenzhen, China). Sequence accuracy of these peptides was confirmed by mass spectrometry and high performance liquid chromatography. Peptide purity was above 95% and the residual endotoxin was measured below 0.025 EU/ml. All of the peptides, provided as lyophilized powders, were dissolved in endotoxin-free water (Sigma) at 20mg/ml and stored at -70°C.
The antigenicity of linear peptides and MAPs {#sec005}
--------------------------------------------
For examination of the antigenicity of linear peptides, enzyme linked immunosorbent assay (ELISA) was performed as described by Elshabrawy *et al*. \[[@pone.0136888.ref022], [@pone.0136888.ref023]\]. Briefly, 96-well plates were coated with 5 μg/ml monoclonal antibody (mAb) against PGN (AbD Serotec, MCA5792, clone No.11-232.3, isotype: mouse IgG3) or mAb against Lipoteichoic acid (LTA, Thermo scientific, MA1-7401, clone No. G35C, isotype: mouse IgG1) at 4°C for overnight. Biotin-labeled SP27, SP27' or L2 as a control peptide (sequence: Biotin-HSGHWDFRQWWQPSGG) was added to the wells, followed by detection with HRP-labeled streptavidin (Jackson ImmunoResearch, 1/10000 dilution). For the examination of MAP27, 5 μg/ml anti-PGN mAb or anti-*S*. *aureus* polyclonal antibodies (stored at our laboratory, 1/800 dilution) were added to the wells coated with 80 μg/ml MAP27 or MAPctrl, and incubated at 37°C for 40 min. HRP-labeled goat anti-mouse and goat anti-rabbit IgG (Jackson ImmunoResearch, 1/10000 dilution) were then added (37°C for 40 min), respectively. Wells were washed three times with 0.1% PBS/Tween 20, followed by addition of the substrate solution containing TMB and H~2~O~2~ for color reaction.
Preparation of bacteria {#sec006}
-----------------------
*S*. *aureus* (ATCC 25923) were purchased from Wenzhou Kont Biology and Technology. The bacteria were grown in tryptic soy broth at 37°C with 250 rpm shaking for overnight. After centrifugation at 6000 rpm for 10 min, the pellet was resuspended and subsequently washed with sterile phosphate-buffer saline (PBS) for two times. The pellet was then diluted with PBS to an appropriate cell concentration as determined by spectrophotometry at 600 nm.
Immunization and ELISA for specific antibodies {#sec007}
----------------------------------------------
The protocol of vaccination was based on our previous study with slight modifications \[[@pone.0136888.ref021]\]. Briefly, female BALB/c mice were divided randomly into three groups. The first two groups were injected subcutaneously with 100 μg MAP27 and 100 μg MAPctrl for five times at 2-week intervals, respectively. The first immunization was performed in Freund's complete adjuvant (Sigma) and the subsequent booster immunizations were administered in Freund's incomplete adjuvant (Sigma). The third group of mice, without any peptide immunization, was breed at the same time and used as the blank control. After 14 days of the fifth MAP immunization, all of the three mice groups (MAP27-immunized, MAPctrl-immunized, and blank control mice) were finally immunized with heat-inactivated *S*. *aureus* (2×10^7^ CFU/mouse in sterile PBS without any adjuvant) via intraperitoneal injection. Serum samples were collected through retro-orbital bleeding 7 days after each immunization and frozen at -70°C for antibody evaluations.
Antibody titers and specificity were measured by ELISA as described previously \[[@pone.0136888.ref021]\]. Briefly, MAP27 (80 μg/ml), sonicated fragments of *S*. *aureus* (OD~600nm~ = 0.5, 50 μl/well) and PGN (10 μg/ml) were coated in 96-well plates at 4°C for overnight, respectively. MAP27- and PGN-coated plates were blocked with 0.25% casein, whereas *S*. *aureus* fragments-coated plate were blocked with blocking buffer containing 1% guinea pig serum to avoid binding of staphylococcus protein A to IgG. The serial diluted antiserum was subsequently added as the primary antibodies, followed by incubation with HRP-conjugated goat anti-mouse IgG as the secondary antibodies. The titer of antibody was defined as the highest dilution that gives more than twice the absorbance value of the blank control.
Mice challenged with lethal live *S*. *aureus* {#sec008}
----------------------------------------------
Five days after the final boost, all of the mice were infected with *S*. *aureus* (ATCC 25923) at 5×10^8^ CFU/mouse via tail vein injection. The mortality was monitored for 7 days (168 hours) post challenge.
Measurement of bacterial burden in organs {#sec009}
-----------------------------------------
To determine the bacterial loads in organs, all of the three groups of mice were infected with *S*. *aureus* at 2×10^7^ CFU/mouse via tail vein. Three days after infection, mice were sacrificed by CO~2~ asphyxiation. The unilateral kidney and lung were aseptically removed and homogenized with 2 ml sterile PBS. The homogenates were then plated on agar media in a 10-fold serial dilution and incubated at 37°C for 18--24 h. The colony forming units (CFU) was determined.
Levels of cytokines in organs post *S*. *aureus* systemic infection {#sec010}
-------------------------------------------------------------------
Mice were euthanized by CO~2~ asphyxiation three days post infection. To measure the local inflammatory response, we centrifuged the homogenized lysates from spleen and lung. Supernatant were collected, added with protease inhibitors (Cell Signaling Technology) and analyzed by ELISA (Biolegend) according to the manufacturer's instructions.
Intracellular cytokine analysis {#sec011}
-------------------------------
Cytokines produced by CD3^+^ T cells were detected by flow cytometry based on a modified method as described by Chen *et al*. \[[@pone.0136888.ref024]\] and Bhattacharya *et al*. \[[@pone.0136888.ref025]\]. Briefly, the splenocytes were harvested three days post infection with *S*. *aureus*. 1--2×10^6^ lymphocytes were incubated with cell stimulation cocktail (containing phorbol 12-myristate 13-acetate, PMA and ionomycin, eBioscience) and brefeldin A (BFA, eBioscience) for 6 h. Cells were then washed, stained with fluorescein-labeled antibodies (FITC labeled anti-CD3 mAb, Cat No. 11--0031, clone:145-2C11; APC labeled anti-CD4 mAb, Cat No.17-0042, clone:RM4-5; eBioscience). Following the fixation and permeabilization, PE-conjugated anti-mouse interferon gamma (IFN-γ) mAb (Cat No.12-7311, clone: XMG1.2, eBioscience) or interleukin 17A (IL-17A) mAb (Cat No. 12--7177, clone: eBio17B7, eBioscience) was used to measure the intracellular cytokine production in CD4^+^ T cells or CD4^-^ T cells, respectively. Data were acquired with a FACS Calibur (BD, Biosciences, San Jose, CA, USA) and analyzed using FCS Express software (De Novo Software, Los Angeles, CA USA). Isotype controls were included in each staining.
*In vitro* stimulation of splenocytes {#sec012}
-------------------------------------
Five days after the last heat-inactivated *S*. *aureus* boost, splenocytes from MAP27-immunized, MAPctrl-immunized and blank mice were obtained by squeezing followed by filtering through a stainless mesh (size, 70 μm) in RPMI1640 medium. Erythrocytes were lysed with RBC lysis buffer (eBioscience). After being washed two times with RPMI1640, cells were resuspended with complete medium (RPMI1640 containing 10% fatal calf serum (Gibico), 100 U/ml of penicillin G and 100 μg/ml of streptomycin) in 96-well U-bottom culture plates. The splenocytes (2 ×10^5^ cells/well) were incubated at 37°C with 100 μg/ml MAP27 for 24 h or 2×10^4^ CFU/well heat-inactivated *S*. *aureus* for 72 h. The supernatants were collected after the incubation and stored at -70°C. The amount of interleukin 2 (IL-2), INF-γ, IL-17A/F and interleukin 4 (IL-4) was determined by ELISA (Biolegend) according to the manufacturer's instructions. All samples were analyzed in triplicate.
ELISPOT assays {#sec013}
--------------
The IFN-γ and IL-17A spot ELISA assays were performed using a kit according to the manufacture's instruction (eBioscience). Briefly, ELISPOT plates (Millipore) were coated with capture antibodies, and incubated at 4°C overnight. The plates were blocked with complete medium at room temperature for 1 h. Splenocytes were harvested and plated at 4×10^5^/well, stimulated with 10 μg/ml MAP27 plus 1μg/ml anti-mouse CD28 antibody or 2×10^5^ CFU/well heat-killed *S*. *aureus* at 37°C for 24 h. Cells and medium were decant from plates, and the plates were washed for three times with washing buffer. Biotinylated detection antibodies and HRP-conjugated avidin were added subsequently. The plates were extensively washed, and developed with substrate solution containing AEC (Sigma) and H~2~O~2~. The reaction was terminated with H~2~O. After drying, the plates were read using CTL analyzer (CTL S5 micro, Cellular Technology Ltd.).
Statistical methods {#sec014}
-------------------
Results were expressed as means ± SEM. Statistical analyses of data to compare the different groups were performed using SPSS version 16.0 software. Data for specificity of antibody, cytokine production and ELISPOT were analyzed using either one-way ANOVA or Student's t-test. Data for flow cytometry and bacterial load were analyzed by Mann-Whitney U of nonparametric Test. For comparison of survival in murine lethal challenge model, individual experiments were analyzed using log-rank test statistic (Mantel-Cox test) from the Prism software (Prism for Windows, version 5.01, GraphPad Software). Probability (*P*) values \<0.05 were considered significant. 4 to 9 mice were used in each experiment and each experiment was repeated two or three times with consistent results.
Results {#sec015}
=======
Both linear peptides and tetra-branched, multiple antigenic peptide MAP27 bind to antibody specifically {#sec016}
-------------------------------------------------------------------------------------------------------
Both of the two linear peptides, SP27 (Biotin-[*SA*]{.ul}SPHHHSRLRSES[*GG*]{.ul}) and SP27' (Biotin-SPHHHSRLRSES[*SAGG*]{.ul}), contain the same core sequence SPHHHSRLRSES, flanked by 2--4 amino acids at N- and C- terminus ([Fig 1A](#pone.0136888.g001){ref-type="fig"}). The core sequence is derived from one of the positive phage clones we previously identified by using an anti-PGN mAb as the target \[[@pone.0136888.ref021]\]. As shown by ELISA, both SP27 and SP27' can bind to anti-PGN in a dose-dependent manner. In contrast, a non-specific peptide L2 (Biotin-HSGHWDFRQWWQPSGG) did not bind to the anti-PGN mAb ([Fig 1B](#pone.0136888.g001){ref-type="fig"}). Moreover, neither SP27 nor SP27' could bind to an unrelated anti- Lipoteichoic acid (LTA) mAb ([Fig 1C](#pone.0136888.g001){ref-type="fig"}). These results indicate that the core sequence (SPHHHSRLRSES), but not the flanking amino acids (SA, GG in SP27, or SAGG in SP27') might mimic the epitope of PGN.
![Both SP27 and MAP27 bind to anti-PGN mAb specifically.\
(A) The sequences of SP27 (Biotin-[*SA*]{.ul}SPHHHSRLRSES[*GG*]{.ul}) and SP27' (Biotin-SPHHHSRLRSES[*SAGG*]{.ul}). Underlined letters represent nonspecific flanking amino acids. (B) Both SP27 and SP27' bind to anti-PGN mAb in a dose-dependent manner. (C) Both SP27 and SP27' (100 μg/ml) specifically bind to anti-PGN mAb, but not to anti-LTA mAb. For ELISA assays in panel B and C, anti-PGN mAb or anti-LTA mAb was used to coat the wells at a concentration of 5 μg/ml, SP27, SP27' or non-specific peptide L2 (in panel B, sequence: Biotin-HSGHWDFRQWWQPSGG) was then added at indicated concentrations and incubated at 37°C for 40 min, followed by detection with HRP-labeled streptavidin. (D) The structure diagram of MAP27. MAP27 was synthesized in a tetra-branched form that contains four copies of a sequence (SASPHHHSRLRSESGG) that mimics PGN epitope. (E) MAP27 binds to anti-PGN mAb specifically. (F) MAP27 binds to anti-*S*. *aureus* polyclonal antibodies specifically. For ELISA assays in panel E and F, MAP27 or MAPctrl was used to coat the wells of a microplate. Anti-PGN mAb or anti-*S*. *aureus* polyclonal antibodies was added, followed by detection with HRP-labeled antibodies. The absorbance was measured at OD~450nm~. The results are shown as means ±SEM. \* P\<0.05, \*\* P\<0.01, \*\*\* P\<0.001.](pone.0136888.g001){#pone.0136888.g001}
Compared with a linear peptide, a multiple antigenic peptide (MAP) has more advantages: it is more stable upon enzymatic degradation; it has bigger size that is sufficient for immunization without crosslinking with any carrier protein; moreover, MAP enhances molecular recognition by immune cells and induces stronger immune responses \[[@pone.0136888.ref026], [@pone.0136888.ref027]\]. Therefore, MAP has been used as an immunogen for developing experimental vaccines against various pathogens \[[@pone.0136888.ref028]--[@pone.0136888.ref032]\]. Based on this fact, we synthesized a MAP, named MAP27 using standard Fmoc chemistry. MAP27 contains four copies of SP27, with C-terminus of each peptide linked with the non-immunogenic, lysine-based dendritic scaffold ([Fig 1D](#pone.0136888.g001){ref-type="fig"}). The predicted molecular weight of MAP27 is 7137.7, big enough for immunization according to our previous work \[[@pone.0136888.ref021]\]. Compared with the attaching backbone (MAPctrl), MAP27 specifically bound to anti-PGN mAb ([Fig 1E](#pone.0136888.g001){ref-type="fig"}) and anti-*S*. *aureus* polyclonal antibodies ([Fig 1F](#pone.0136888.g001){ref-type="fig"}). These results indicate that a similar antibody-binding epitope as PGN was remained in MAP27.
Last boost with heat-inactivated bacterium produced low titer antibody against *S*. *aureus* in MAP27- immunized mice {#sec017}
---------------------------------------------------------------------------------------------------------------------
We then evaluated the immunogenicity of MAP27. Serum samples from mice immunized with MAP27 or MAPctrl were used for binding assays by indirect ELISA. The titer of antibodies against MAP27 in MAP27-immunized mice reached to 10^4^ after the third time of vaccination, whereas no detectable antibodies were produced in MAPctrl-immunized mice ([Fig 2A](#pone.0136888.g002){ref-type="fig"}). We also found that serum from MAP27-immunized mice weakly reacted with PGN after the fifth MAP immunization ([Fig 2B](#pone.0136888.g002){ref-type="fig"}, at the 1/200 dilution). In contrast, antibodies to *S*. *aureus* were not detected in the same samples (data not shown). As shown in previous studies, boosting with natural antigens after immunization with peptide could significantly enhance the immune response \[[@pone.0136888.ref033], [@pone.0136888.ref034]\]. Therefore, we boosted all of the three groups of mice with heat-inactivated *S*. *aureus* for only one time. Anti-*S*. *aureus* IgG from MAP27-immunized mice could be detected with a dilution of 1/200. Conversely, the serum samples from either MAPctrl-immunized mice (P = 0.016) or blank control mice (P = 0.002) produced no detectable antibodies ([Fig 2C](#pone.0136888.g002){ref-type="fig"}). These results indicate that *S*. *aureus* boosting specifically augmented the immune responses triggered by MAP27 immunization.
{#pone.0136888.g002}
Vaccination with MAP27 protected the mice against *S*. *aureus* lethal-challenge and significantly reduced the bacterial burden in organs {#sec018}
-----------------------------------------------------------------------------------------------------------------------------------------
We then tested whether MAP27 immunization protects the mice against *S*. *aureus* infection. All of the mice were challenged with live *S*. *aureus* at a lethal dose on the fifth day post the last boost of heat-inactivated bacterium. The survival was then monitored. After seven days of lethal challenge, 50% of the MAP27-immunized mice were still alive, whereas, all of the control mice and MAPctrl-immunized mice died on day two (P = 0.000) and day five (P = 0.000), respectively ([Fig 3A](#pone.0136888.g003){ref-type="fig"}). This result indicates that MAP27 vaccination protects the mice against *S*. *aureus* lethal challenge.
{#pone.0136888.g003}
To further determine the effect of MAP27 immunization on bacterial growth in animal organs, all of the mice were sacrificed three days after *S*. *aureus* infection. We found that the number of bacterium in the kidney of MAP27-immunized mice is significantly fewer than that of MAPctrl-immunized mice (P = 0.029) or blank control mice (P = 0.029) ([Fig 3B](#pone.0136888.g003){ref-type="fig"}). Similar results were also observed in lung ([Fig 3C](#pone.0136888.g003){ref-type="fig"}, MAP27 vs MAPctrl, P = 0.016; MAP27 vs blank control, P = 0.008). Taken together, these experiments indicate that vaccination with MAP27 reduced bacteria burden in animal organs.
IFN-γ, IL-17A/F and CCL3 were increased in organs of MAP27-immunized mice after *S*. *aureus* systemic infection {#sec019}
----------------------------------------------------------------------------------------------------------------
As we described above, although MAP27 immunization only induced the production of low titer antibodies against *S*. *aureus* or PGN, it protected the mice against *S*. *aureus* systemic infection. This fact led us to speculate that the protection effect is possibly through the *in vivo* T cell-mediated response. To test this hypothesis, we examined T cell-derived cytokines in spleen and lung from the mice three days post infection. As expected, the levels of IFN-γ and IL-17A/F, cytokines mainly secreted by T cells, were remarkably higher in MAP27-immunized mice than in MAPctrl-immunized mice and blank control mice ([Fig 4](#pone.0136888.g004){ref-type="fig"}). CCL3, a chemokine that activates macrophages and neutrophils \[[@pone.0136888.ref035], [@pone.0136888.ref036]\], as well as enhances IFN-γ production \[[@pone.0136888.ref037]\], was also increased in MAP27-immunized mice ([Fig 4](#pone.0136888.g004){ref-type="fig"}). These results indicate that T cell responses might mediate the protection of MAP27 immunization against the *S*. *aureus* systemic infection.
{#pone.0136888.g004}
The IFN-γ^+^CD3^+^ and IL-17A^+^CD4^+^ T cells were increased in spleen of MAP27-immunized mice post *S*. *aureus* systemic infection {#sec020}
-------------------------------------------------------------------------------------------------------------------------------------
We further characterized cytokine expression profiles of infiltrated T cell population over the course of infection. To enhance the cytokine signals detected by flow cytometry, we conducted the PMA and inomycin treatment. As shown in [Fig 5A and 5C](#pone.0136888.g005){ref-type="fig"}, the percentage of IFN-γ^+^CD3^+^CD4^+^ T cells from MAP27-immunized mice (about 7.18±0.96%) was significantly higher than that from MAPctrl-immunized mice (3.85±0.41%, P = 0.021) or blank control mice (3.57±0.73%, P = 0.03). Similar results were also observed for IFN-γ^+^CD3^+^CD4^-^ T cells ([Fig 5B and 5C](#pone.0136888.g005){ref-type="fig"}, MAP27 vs MAPctrl, P = 0.021; MAP27 vs blank control, P = 0.012). These results suggest that both CD4^+^ and CD4^-^ T cells were stimulated to produce IFN-γ in MAP27-immunized mice post *S*. *aureus* infection.
{#pone.0136888.g005}
Th17 cells have been reported to play an important role in clearance of *S*. *aureus* \[[@pone.0136888.ref014], [@pone.0136888.ref038]\]. Here we also found the percentage of IL-17^+^CD3^+^CD4^+^ T cells from MAP27-immunized mice (about 2.4±0.38%) was higher than that from MAPctrl-immunized mice (about 1.34±0.13%, P = 0.021) and blank control mice (0.31±0.08%, P = 0.012) ([Fig 5D and 5F](#pone.0136888.g005){ref-type="fig"}). The percentage of IL-17^+^CD3^+^CD4^-^ T cells is not statistically different among the three groups ([Fig 5E](#pone.0136888.g005){ref-type="fig"}). Taken together, not only IFN-γ- but also IL-17-producing T cells of MAP27-immunized mice were stimulated effectively upon *S*. *aureus* infection.
*In vitro* MAP27 treatment stimulated the splenocytes from MAP27-immunized mice to produce IFN-γ {#sec021}
------------------------------------------------------------------------------------------------
To further assess whether immunization with MAP27 was able to induce IFN-γ and IL-17 *in vitro*, we harvested splenocytes after five days of the last boost with heat-inactivated bacteria. Cells were incubated with 100 μg/ml MAP27 for 24 h. The supernatant was then collected and cytokine induction was analyzed by ELISA. The splenocytes from MAP27-immunized mice produced significantly more IL-2 and IFN-γ than cells from control mice upon stimulation by MAP27 ([Fig 6A and 6B](#pone.0136888.g006){ref-type="fig"}). In contrast, IL-4, which is mainly produced by Th2 cells, could only be detected very weakly in all three groups ([Fig 6C](#pone.0136888.g006){ref-type="fig"}). This can be explained by the fact that high level of IFN-γ in MAP27-immunized mice might suppress Th2 cell differentiation \[[@pone.0136888.ref039]\]. Unexpectedly, we failed to detect IL-17A/F in supernatant of spleen cells from all the mice (data not shown).
{#pone.0136888.g006}
We then used ELISPOT, a more sensitive assay than ELISA, to further determine whether the T cell-mediated response was induced by MAP27. As shown in [Fig 6D and 6E](#pone.0136888.g006){ref-type="fig"}, the number of IFN-γ-producing cells from MAP27-vaccinated mice was significantly higher than that from MAPctrl-immunized mice (P = 0.001) or blank control mice (P = 0.001). Again, we could not identify IL-17A-producing cells by ELISPOT assay. In light of this, we speculated that immunization with MAP27 might predominantly induce Th1 cell response.
Splenocytes from MAP27-immunized mice specifically recognized *S*. *aureus* and produced IFN-γ and IL-17 {#sec022}
--------------------------------------------------------------------------------------------------------
Our next question was whether *S*. *aureus* as a natural antigen could be recognized by splenocytes from MAP27-immunized mice. Splenocytes isolated from three groups of mice were incubated with heat-inactivated *S*. *aureus* for 72 h, and cytokines in the culture supernatant were analyzed by ELISA. Two classical cytokines produced by T cells, IL-2 and IFN-γ, were significantly produced in splenocytes from MAP27-immunized mice, but not MAPctrl-immunized mice or blank control mice ([Fig 7A and 7B](#pone.0136888.g007){ref-type="fig"}). Interestingly, IL-17A/F, a cytokine mainly secreted by Th17 cells, is also significantly induced in splenocytes from MAP27-immunized mice ([Fig 7C](#pone.0136888.g007){ref-type="fig"}). Similar to the *in vitro* assays with MAP27 stimulation ([Fig 6C](#pone.0136888.g006){ref-type="fig"}), IL-4 was not detectable in any group ([Fig 7D](#pone.0136888.g007){ref-type="fig"}). We next performed ELISPOT assay to quantitate IFN-γ- or IL-17A-producing cells after *S*. *aureus* stimulation. As shown in [Fig 7E--7H](#pone.0136888.g007){ref-type="fig"}, the number of IFN-γ- or IL-17A-producing cells from MAP27-immunized mice was significantly higher than that from the control groups (IFN-γ-producing cells: MAP27 vs MAPctrl, P = 0.0126; MAP27 vs blank control, P = 0.0041; IL-17-producing cells: MAP27 vs MAPctrl, P = 0.0034; MAP27 vs blank control, P = 0.0039). Collectively, these results indicate that T cells, especially IFN-γ- and IL-17-producing T cells in MAP27-immunized mice, recognized effectively the nature antigen *S*. *aureus in vitro*.
{#pone.0136888.g007}
Discussion {#sec023}
==========
To date, the failure to develop an effective vaccine against *S*. *aureus* is partly due to the high complexity of the numerous virulence factors that are expressed at different infection stages. Considering this, Lin and colleagues suggested that immunogens might not be restricted to microbial virulence factors, but any target antigen that effectively induces T cell immune responses could be selected \[[@pone.0136888.ref016]\]. Here, we used PGN, a conserved cell wall component of *S*. *aureus* as the target, and designed a *de novo* synthetic multiple antigenic peptide (MAP27) for mimicking a PGN epitope. The immunization outcome could be affected by various route of immunization. For example, vaccination of antigens via eyes \[[@pone.0136888.ref040], [@pone.0136888.ref041]\] or oral route \[[@pone.0136888.ref042], [@pone.0136888.ref043]\] has been shown to induce specific tolerance in BALB/c mice. In this study, we found that subcutaneous immunization, with peptides emulsified with Freund's adjuvant, effectively induced immune response and protected the mice against *S*. *aureus* systemic infection. It would be very interesting to test whether other routes of immunization, such as via mucosal immunization, could further enhance the protective effect.
As shown in [Fig 2B](#pone.0136888.g002){ref-type="fig"}, immunization with MAP27 alone could only weakly induce IgG to bind to PGN. This can be explained by the fact that the peptide only mimics a single epitope of PGN, and possibly the peptide mimicking is not completely identical to the natural antigen. Moreover, as an artificial antigen, MAP27 probably has weaker immunogenicity compared to most of the protein immunogens. Since previous findings have showed that boosting with natural antigen after peptide immunization could induce a robust recall response \[[@pone.0136888.ref033], [@pone.0136888.ref034]\], we modified our immunization strategy by boosting the mice with heat-inactivated *S*. *aureus* after MAP immunization. Our results demonstrated that boosting with natural antigen improved the binding of the sera to the bacteria ([Fig 2C](#pone.0136888.g002){ref-type="fig"}). In contrast, the same boosting step had no effect on the control groups (Figs [2C](#pone.0136888.g002){ref-type="fig"} and [3A](#pone.0136888.g003){ref-type="fig"}). This result is consistent with the previous finding that vaccination with heat-killed *S*. *aureus* before infection does not affect bacterium outcome \[[@pone.0136888.ref044]\].
Although the titer of sera against PGN or *S*. *aureus* is low, MAP27-immunization could still protect the mice from *S*.*aureus* systemic infection ([Fig 3](#pone.0136888.g003){ref-type="fig"}). Previous studies have demonstrated that IL-17 and Th17 are critical for protection from *S*.*aureus* infection \[[@pone.0136888.ref009], [@pone.0136888.ref011], [@pone.0136888.ref014]\]. Moreover, IFN-γ is universally considered to improve the survival of staphylococcemia mice by activating marcrophages and neutrophils to kill bacteria \[[@pone.0136888.ref045]\]. M1-polarized macrophages stimulated with IFN-γ could also inhibit the formation of *S*. *aureus* biofilm \[[@pone.0136888.ref046]\]. Based on these facts, we hypothesized that the protection is possibly through the activation of T cell-mediated responses. Several evidences to support this hypothesis were presented in this study: 1) Levels of IFN-γ and IL-17A/F were significantly increased in the early stage of bloodstream infection ([Fig 4](#pone.0136888.g004){ref-type="fig"}); 2) Both IFN-γ^+^CD3^+^CD4^+^ Th1 cells and IL-17^+^CD3^+^CD4^+^ Th17 cells significantly increased in MAP27-immunized mice after infection ([Fig 5](#pone.0136888.g005){ref-type="fig"}); 3) Our *in vitro* assay showed that splenocytes from MAP27-immunized mice specifically recognized *S*. *aureus* and produced IFN-γ and IL-17 ([Fig 7](#pone.0136888.g007){ref-type="fig"}). Interestingly, we found that IFN-γ^+^CD3^+^CD4^-^ T cells also increased in MAP27-immunized mice post infection ([Fig 5B and 5C](#pone.0136888.g005){ref-type="fig"}). We speculated that MAP27 immunization might induce different subsets of T cells to secrete IFN-γ post infection. For example, CD8^+^ T and γδ T cells have been reported to secrete IFN-γ and protect mice from *S*. *aureus* systemic infection \[[@pone.0136888.ref017], [@pone.0136888.ref047], [@pone.0136888.ref048]\].
Different from heat-inactivated *S*. *aureus* stimulation ([Fig 7](#pone.0136888.g007){ref-type="fig"}), MAP27 failed to induce IL-17A/F in splenocytes (data not shown), even though it induced high level of IFN-γ *in vitro* ([Fig 6B](#pone.0136888.g006){ref-type="fig"}). Again, this result can be explained by the fact of the inhibitory role of IFN-γ in Th17 differentiation \[[@pone.0136888.ref049]\]. We speculated that the artificial MAP antigen and the natural antigen (*S*. *aureus*) might trigger different responses. However, the exact mechanism is still unclear and needs to be elucidated. A previous study suggested that MAP can be internalized by antigen presenting cells (APCs) and presented as an epitope on cells \[[@pone.0136888.ref050]\]. It would be very interesting to investigate how APCs recognize this peptide to elicit T cell responses.
In summary, our results indicate that immunization with PGN mimicking peptide, MAP27, could protect mice against *S*. *aureus* systemic infection. This protection might be related with the stimulation of IFN-γ- and IL-17-producing T cells that specifically recognized *S*.*aureus*.
Supporting Information {#sec024}
======================
###### MAP27 immunization decreases the bacterial burden in organs.
Mice were immunized with MAP27 or MAPctrl for five times at a two-week interval without bacterial boost (A, C), or with bacterial boost (B, D). All the mice were infected with *S*. *aureus* via tail vein post five days of the last immunization. The bacterial numbers in organs were measured after three days of infection. Bacterial burden in kidney (A, B) and in lung (C, D) were measured. (\* P\<0.05; \*\* P\<0.01) (n = 4-11/group).
(TIF)
######
Click here for additional data file.
We thank Da-Ming Zuo for his kind advice on design and technical assistance. We also thank Li-Chang Liu for handling of experimental animals.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: BYL NF. Performed the experiments: XYW ZXH YGC XL KW. Analyzed the data: XYW ZXH. Contributed reagents/materials/analysis tools: PZ. Wrote the paper: BYL NF.
[^3]: Current Address: Medical Laboratory, Jiangxi Provincial People's Hospital, Nanchang, Jiangxi, People's Republic of China
|
{
"pile_set_name": "PubMed Central"
}
|
On the DMF strategy, as I do not have my Infinite Whelpling maxed to 25 yet, I substituted a Spawn of Onyxia rare (4). I used the same skills as the Infinite with one exception: Lift-Off in place of Borrowed Time. Same strategy as posted and Jeremy was defeated.
Three Peck's won't kill Tiptoe, but three* Pecks and an Adrenaline Rush will. The Adrenalin Rush has the added bonus of ensuring that even when you lose the Flying racial, you will still be faster than Tiptoe which means you can kill them before they kill you. If Tiptoe never Tidal Waves, then you can be hit with 3 Water Jets. This has happened to me; the Gryphon has enough to survive them. The only way Tiptoe should be able to kill the Gryphon is if it only uses Water Jet *and* it gets a crit. So, it could happen, but its pretty unlikely.
* - If Tiptoe heals himself, then, of course, you will need four Pecks but since the Peck and heal cancel each other out and you don't take any damage that turn, we can basically ignore it. That is, though, why the Adrenaline Rush comes after the first Peck to ensure the buff is still up for the fourth Peck if Tiptoe heals.
I've been able to trim up Sludgy a bit in the Thundering Spirit fight. If you use the Sen'jin Fetish with Flame Breath and Rot, sludgy will always go down in 4 rounds of Flame Breath as it is a 100% hit chance and always has enough damage to kill Sludgy. Rot is for the unlikely scenario that your undead round extends into the first round of Darnak. 3 possible outcomes:
Never seen this, but on the extremely odd chance that the fetish doesn't die:Swap in Sen'jin FetishFlame BreathFlame BreathFlame BreathFlame Breath (Sludgy dies)Swap to Carry Pet(Darnak's first round)Swap to Rabbit (Darnak burrows)
The nice thing here, is sludgy always dies in 4 damage rounds (zepp and dragonling can both sometimes do this, but not guaranteed) and it is very RNG resistant. Plus if you get lucky, you get a Rot on Darnak for added damage from your remaining pet (I usually use a rabbit).
Sorry I've been on a brief hiatus from pet battling so hadn't made any changes lately. I've fixed the various location errors (thanks for pointing them out). I added Stencil's improvement to the Flowing Pandaren Spirit fight since that's very straightforward. The other changes people have suggested look good (even if it means retiring the Major Payne team I've used for so long) - I need to level up my Sen'Jin fetish, that looks like it's even smoother than the Scourged Whelpling since it has a 0% miss (vs Tail Sweep's 5% miss chance). I just need to sit down and get the work done (though Stencil's done the work for his suggestions, so for me it's just making the edits).
Keep up the great work, all, this is a community guide and anything you can come up with that can improve the guide (while keeping to the goals) is appreciated.
Rendigar wrote:You can go thru any Blasted Lands portal and fly over to fight Lydia Accoste in Deadwind Pass (40,76), and also go fight Trixxy in Winterspring (65, 64) as they both also give a Sack of Pet Supplies (any of which can contain Flawless-type Battle-Stones). They do not give good XP for your pets, assuming you use 25’s and a carry pet, but the sacks may be worth the effort.
I would not list a strategy for them because any pair of 20-somethings can beat them - if you have gotten this far I know you can beat them with 2 pets if you want the XP from the fight and not just the bag.
No experience for carry pet if it doesn't participate. For Thundering Pandaren Spirit, you have "Note: your carry pet doesn't need to participate, as it will be the only one left alive at the end of the fight and will thus receive XP", but that did not happen for me.
Edit: You're right, carry pet DOES get XP if other 2 are dead, even if carry pet doesn't participate. Did this on Thundering Spirit, with Rapana Whelk and Darkmoon Zep exploding at the end; my carried Zandalari Toenibbler got 4813 XP. I must not have been paying attention earlier.
Last edited by Gilmar on January 26th, 2014, 2:05 pm, edited 1 time in total.
I started to make a guide just like this! I was going to post it... then found its all been done before... oh well. Here is what I found. Maybe it will help with some alternative ways of doing it.Guide to Leveling Pets Using Tamers of Pandaria, Northrend and Cataclysm (5.4)
If a pet you are trying to level needs to do some damage or survive or do some damage as listed in this guide; you can simply use Mechanical Pandaren Dragonling or some other pet with explode or something like that to get it a few levels first; either with wild pets or some of the other trainers that don’t require that they do any thing. All versions of pets used should be rare quality in order to do well or in many cases to function right. Remember when doing this, dead pets earn no experience, and surviving pets that entered combat will gain all the experience from the battle. So if the pet your are leveling is the only pet to survive the battle, he will gain all the experience even if he didn’t enter combat. Also, if a pet that your leveling entered combat and survived, even with a max level pet which survived, all the experience will go to the low level pet. This switching technique works if you bring in the low level after a pet dies then swapping back out, but will NOT work if you begin the battle with the low level pet and swap him out first turn.
Thundering Pandaren Spirit:Snail - Ooze Touch, Acid Goo, DiveDarkmoon Zeppelin - Missile, Explode, DecoyPet to level does not need to do any thing if things go well. Kill start with snail. Use dive first turn to avoid rupture, if it misses you can restart with no penalty. Then acid goo and ooze touch twice to kill first pet. Damage 2nd pet using ooze touch first then acid goo. This is where your snail usually dies and should be able to finish off with Darkmoon Zeppelin with 3 Missiles. On the rat, open with missile, decoy, then missile till it has 618 hp or less and explode.
Whispering Pandaren Spirit: Celestial Dragon - Flamethrower, Ancient Blessing, MoonfireInfected Squirrel - Stampede, Creeping Fungus, Corpse ExplosionPet to lvl may need to survive a little bit of damage from cyclone. Celestial Dragon or Nether Faerie Dragon can down the first 2 pets by himself. Its fairly easy to figure out, though it may be a little tricky because the stun from the moth is a pain, but with a little luck he can even do a little damage to the third, if he only manages to do some damage to the dragon, all is not necessarily lost. Start with dragon, use moonfire, ancient blessing, flamethrower, flamethrower, moonfire moth is usually dead here. Then use ancient blessing, when he brings up Whispertail, fallowed by flamethrower, flamethrower, moonfire then this is where the second pet might die. Just flamethrower and moonfire till your dragon is dead. If you were able to kill the first 2 pets with the dragon, bring out the low level pet, then swap to the Infected Squirrel. Use creeping fungus, then Stampede till the elemental is dead, or if your squirrel dies first and the undead breed bonus triggers use corpse explosion. On this one, I use the swap technique if the first 2 enemy pets die from your dragon because if the elemental dies at the same time the squirrel dies, the battle will end before the undead bonus wears off and counts as having survived the fight. Thus your low level pet would not otherwise gain any experience. However if things don’t go so well and your dragon dies before it kills the other, and the squirrel needs to finish it off, use stamped to do it then when it ends and elemental is up use Creeping Fungus and stampede again if you can. If you can tell your Infected Squirrel is going to die soon and you are not stampeding, use corpse explosion. Since the elemental goes first and kills him, then the damned effect triggers before Corpse Explosion is cast, the squirrel will still have another turn.
Hyuna of the Shrines: Sunreaver Micro-Sentry - Laser, Extra Plating, Call LightningDarkmoon Zeppelin - Missile, Explode, Decoypet to level doesn’t need to do any thing. Start with Sunreaver Micro-Sentry, Laser then use Extra Plating. Then Laser till the bug is dead. When snake comes up use Laser, then use extra plating, Call Lightning, Laser then next turn Sentry usually dies if he didn’t already. Bring in Zeppelin and use Decoy. Its perfect for the turtle because his only attack is headbutt which has a 4 round cool down and decoy has 2 charges with an 8 round cool down. just remember to use explode when the turtle is down to 618 hp or less and make sure he isn’t about to use a heal.
Farmer Nishi:Water Skimmer - Water Jet, Cleansing Rain, Pump (281+ speed works best, may need to change the order you use abilities if slower)Darkmoon Zeppelin - Missile, Explode, DecoyPet to level will need to be able to survive at least about 167 points of elemental damage, so more if its a mechanical. Start with Water Skimmer, and use Pump first, then Cleansing Rain, then Water Jet until the first elemental is dead. Then for the next elemental start with a pump, then pump and cleansing rain while he is submerged, then Water Jet him till dead, but don’t kill him. On the beast use the pump you have stored, then he should die after the burrow, if not just water jet till dead. However unlikely make sure not to kill the last pet with your skimmer. Use Decoy on Zeppelin then missile till he is about 600 hp or lower, then use explode, make sure to keep track on Burrow it has a 4 round cool down.
Mo’ruk: Darkmoon Zeppelin - Missile, Explode, DecoyCelestial Dragon (OR Nether Faerie Dragon) - Flamethrower, Ancient Blessing, MoonfirePet to level does not need to do any thing. Start with Zeppelin, use missile, then decoy, then missile till dead. As soon as moth comes up swap to Celestial Dragon, then use moonfire, ancient blessing, then flamethrower till moth is dead, assuming you don’t get stunned from moth dust. When the turtle comes up use moonfire if it can survive long enough to do so then flamethrower till your dragon dead. Use Zeppelin to bring the turtle to about 600hp or lower and use explode.
Flowing Pandaren Spirit:Moth - Wild Winds (or Slicing Wind), Cocoon Strike, Moth DustMechanical Pandaren Dragonling(made rare) - Breath, Bombing Run, Explodepet to level needs to be able to survive possibly a lot of aquatic damage. Start with moth, use moth dust, if it misses, you can forfeit and try again. Then slicing winds, cocoon strike, slicing wind, slicing wind and first pet is usually dead here. If moth dust stuns he will swap to second pet sooner, just kill with slicing wind and moth dust, and save cocoon strike for when he switches back to the first one to use agains the last tick on whirlpool. Deal some damage with moth to last pet, make sure not to kill. Bring out zeppelin, use missile to bring down to 618 hp or less and explode. Straight forward fight, just a lot of tidal wave attacks for your low level pet to survive.
Wastewalker Shu:Moth - Wild Winds (or Slicing Wind), Cocoon Strike, Moth DustMechanical Pandaren Dragonling(made rare) - Breath, doesn’t matter, Explode (or Zeppelin)Pet to lever will need to survive a small bit of damage from quake. Start with Moth, use moth dust, and slicing winds to damage, and save cocoon strike for the last tick of whirlpool. If crab is low health, you may wish to save moth dust for the next pet. Most of the time I can kill the first 2 pets with moth. Then just use Mechanical Dragon to bring down the porcupine to 560 health or lower and explode. The only reason I use the dragonling and not the zeppelin here is because if you should get stuck on the elemental when the moth dies, the zeppelin only has a mechanical attack which doesn’t help agains an elemental.
Seeker Zushi:Electrified Razortooth - Rip, Blood in the Water, DevourDarkmoon Zeppelin - Missile, Explode, DecoyPet to level will need to be able to survive a small bit of elemental damage. Start with Electrified Razortooth, use rip, blood in the water, then devour. After devour is done the elemental should be dead. Then the snail comes up use rip then swap to Zeppelin and use Decoy, then swap back to razortooth. Use rip, devour. Snail usually dies here, use rip, blood in the water rip then pass till razortooth is dead. Use Zeppelin explode for the win. If razortooth dies just before the snail does because he didn’t get healed from devour all is not lost, zeppelin can still finish off the snail and kill the aquatic with the help of his decoy.
Burning Pandaren Spirit:Bonkers - Jab, Doge, Tornado PunchSnail - Ooze Touch, Acid Goo, DivePet to level will need to be able to survive a little damage from Cyclone, which can do about 120 damage each round for 5 rounds, but only 35% hit chance. Start with Bonkers and use Tornado Punch (just restart if it misses), doge, then jab till the dragon is dead. Elemental will become active, use jab and tornado punch until bonkers is dead. Bring out your low level pet, then swap to your snail. The sail is good against both the elemental and fire fly.
Courageous Yon:Chrominius - Arcane Explosion, Ancient Blessing, (Surge of Power or Ravage)Darkmoon Zeppelin - Missile, Explode, DecoyPet to level does nothing. Use Chrominius’s arcane explosion to kill the chicken, use ancient blessing when its not on cool down starting after the first lift-off lands. When the Rabbit comes up use surge of power, then chrominus dies here. Use Zeppelin’s missile to finish off the rabbit then bring the ram down to 618 hp or lower and explode. If you use Ravage instead of surge of power, just use it whenever it would kill a pet so it will heal him. In fact, that way you can kill the first 2 pets completely and put the last one into range of a zeppelin explosion.
Aki the Chosen:Curious Oracle Hatchling - Water Jet, doesn’t matter, Dreadful BreathChrominius - Arcane Explosion, Ancient Blessing, RavagePet to level does nothing. Start with Oracle Hatchling, use Dreadful Breath till the turn just before he dies (around 500 or 600 hp left) and use water jet twice. Then use dreadful breath till either the oracle dies or the dragon in the back line dies. When oracle dies, bring out low level pet then swap to Chrominius. Use arcane explosion till the dragon in the back line dies if he isn’t already, then ancient blessing, ravage, and the grasshopper should die here if he isn’t already dead. Other wise IF both grasshopper and dragon is dead before the oracle hatchling dies, then just water jet or ravage to kill the otter, just remember to swap to low lvl pet at some time to make sure he gets the experience.
While your doing the Northrend ones, be sure to stop by the spawn locations for the Unborn Val’kyr if you don’t already have one.
Major Payne:Darkmoon Zeppelin - Missile, Explode, DecoyInfested Bear Cub - Diseased Bite, doesn’t matter, Corpse Explosion(or instead of Infested Bear Cub you could probably use Gulp Froglet)Pet to level doesn’t need to do any thing most of the time. This is one of my least favorite battles. I’ve found this actually works best with Zeppelin first then swap to Phoenix Hatchling for the speed stat, but I realize that pet can be hard to come by. Start with Zeppelin and use decoy, the bear will either use bash or rampage first. I don’t know what determines it, seems random. If he uses bash I recommend forfeiting till he doesn’t. Use Decoy then just Missile till dead. If one missile misses, sadly your pretty hosed. Beakmaster X-225 comes out damage it as much as you can, I’ve been able to successfully down him past his first health by using decoy to survive long enough to get him in range of explode. When zeppelin dies, use infested bear cub to finish him off, then bloom becomes active. Use diseased bite till he becomes undead then use corpse explosion. Very difficult fight. Good luck, you will literally need it.
Nearly Headless Jacob:Swamp Croaker - Tongue Lash, Swarm of Flies, BubbleDarkmoon Zeppelin - Missile, Explode, DecoyPet to level does nothing. Start with Swamp Croaker, use swarm of flies and lash till the first pet is dead. He will make you swap pets with Death grip, just swap back. All enemy pets are undead, so they will come back to life for 1 round, when the first one does that use bubble to block the call darkness from the second one.
Gutretch:Clockwork Gnome - Mettle Fist, Repair, Build TurretDarkmoon Zeppelin - Missile, Explode, DecoyPet to level doesn’t need to do any thing. Start with Gnome, use Build Turret, Mettle Fist, Repair, Mettle Fist.(fist pet dead) Build Turret, then Mettle Fist till gnome is dead. When he is dead, bring out Zeppelin and missile the rat until its in range of explosion, which is 618hp
Orkut Dragonwaste:Scourged Whelpling - doesn’t matter, Call Darkness, Dreadful BreathDarkmoon Zeppelin - Missile, Explode, DecoyPet to level does nothing. Start with Scourged Whelpling and use Dreadful Breath then when its done Call Darkness. Use Dreadful Breath again until your whelpling dies. Then Zeppelin uses Missile till the second pet dies once then on the turn it comes back to life use decoy. When the third pet comes up use Missile till its dead, and on the turn it comes back to life use explode to ensure the pet to level is the only survivor.
Beegle Blastfuse: Nether Faerie Dragon - Arcane Blast, Exchange Life, MoonfireDarkmoon Zeppelin - Missile, Explode, DecoyPet to level does nothing. Start with Nether Dragon. The first bird some times opens with lift off, but it doesn’t matter to me just use arcane blast any way because it triggers its extra damage for the next time even if it misses. When second pet stuns him, just pass turn, and just use arcane blast, it does more damage now than moonfire. However if you get down to 300hp with Nether Dragon, you will need to use Life exchange just to stay alive. Try to get in a moonfire or something on third pet and just damage till dragon is dead. Then you should know what to do by now about pets with explosion, just remember to use decoy just before Ice Tomb lands… or not, probably got lots of health and a mechanical revive to get you through.
Obalis:Darkmoon Zeppelin - Missile, Explode, DecoyCelestial Dragon - Flamethrower, Ancient Blessing, MoonfirePet to level does nothing. Start with Zeppelin, use decoy, then missile to death. When moth comes up switch to your dragon, and use flamethrower, moonfire, ancient blessing, flamethrower. Flamethrower and moon fire till moth is dead, and keep doing that till your dragon is dead from the critter, bring out zeppelin and use to damage critter down to 618 hp or lower and explode.
Goz Banefury:Mechanical Pandaren Dragonling(made rare) - Breath, Thunderbolt, DecoyDarkmoon Zeppelin - Missile, Explode, DecoyPet to level does nothing. This one is waaaaaaaaaaay to unpredictable but to help come up with something that might work for you; start with Mechanical Pandaren Dragonling, use thunderbolt, breath, decoy, breath, breath, thunderbolt, pass for dark rebirth, breath then phoenix may die here. Amythel comes up, use breath till you can use decoy then thunderbolt and the Amythel usually dies here. Spider becomes active, use breath a few times till mechanical dragon is dead, then bring out zeppelin and explode, or missile a few times till spider is down to 618 hit points then do it.
Brok:Flayer Youngling - Triple Snap, Deflection, RampageDarkmoon Zeppelin - Missile, Explode, DecoyPet to level needs to survive about 175 magic damage more if its a flying pet while he is waiting his turn, then a 2 rounds of swarm. Start with Flayer Youngling, use deflection, rampage, deflection, triple snap, triple snap, then first pet should be dead here. Lizard comes out, use rampage, and a triple snap if it survives, then it should be dead next turn or so. Use missile from zeppelin till lizard is dead, then roach comes out. Use missile twice then, decoy. Try to use missile till its about 618 health or lower. Unfortunately the roach will probably use survival, thats why your pet to level will need to be able to survive a lot of damage to land at least 1 hit point of damage.
Bordin Steadyfist:Electrified Razortooth - Rip, Blood in the Water, DevourDarkmoon Zeppelin - Missile, Explode, DecoyPet to level does nothing. Start with Razortooth, use rip blood in the water, devour, first pet dies. critter becomes active, devour continues, then use rip, then devour, critter usually dies after devour finishes. Last elemental becomes active, use rip, and devour, then pass till your pet is dead. I know, self restraint is soooooo hard to come by. Then explode with zeppelin for the win.
When fighting Aki the Chosen, I've been using Zandalari Raptor + Clockwork Gnome + carry pet. Raptor kills the cricket, Stormlash kills the raptor, and Gnome kills Stormlash with a turret. Whiskers comes out, I bring out the carry pet, Whiskers dives. I bring out the Gnome who eats the dive and kills Whiskers.
At the end of the fight, my carry pet gets no XP. No level up, nothing in the log, etc. Anyone else seen this happen? The first time, I thought that maybe I had forgotten to bring out the carry pet, so from then on I was very careful with that trainer. But I've seen it happen twice since then.
I think it is a bnet thing. I saw this 3 times yesterday when I soloed wild teams with my level 24 and still got no EXP. Now on AKI, I do sometimes get no indication of getting EXP like you describe, but my pet has always gotten that EXP regardless (At least they always gain a level or more).
Instead of Water Jeting until the Eternal Dies, switch to carry pet, mothran will dive, switch to Tonk, to be safe S&A then Cannon, but unless his heal hits you for a crit Canon will kill him. The tonk takes the hit, but this usually cuts a couple rounds off, even if he does survive the canon, he will hot survive S&A after you recover.
Hammered wrote:When fighting Aki the Chosen, I've been using Zandalari Raptor + Clockwork Gnome + carry pet. Raptor kills the cricket, Stormlash kills the raptor, and Gnome kills Stormlash with a turret. Whiskers comes out, I bring out the carry pet, Whiskers dives. I bring out the Gnome who eats the dive and kills Whiskers.
At the end of the fight, my carry pet gets no XP. No level up, nothing in the log, etc. Anyone else seen this happen? The first time, I thought that maybe I had forgotten to bring out the carry pet, so from then on I was very careful with that trainer. But I've seen it happen twice since then.
It is a bug with the Aki fight. Your pet should still get XP (if you remembered to swap him in), but the log doesn't show it.
Just wanted to post my experience regarding the listed strategy for Nitun.... burst team 2 is tagged "ideal" against him, implying that it should pretty much always work, but I rarely see it work.
Most of my battles with this team end up with Nitun healing back to high health. Surge of Power only occasionally gets off (usually Chrominius is killed before he gets a chance to use Surge of Power, and I've seen him killed the turn he comes out before he can even get off a howl). If I do manage to get off surge of power, it's not enough to finish him off, and Chrominius never lives long enough to get off a bite and a surge of power, so he then dies to ravage which allows Nitun to heal... after that, the Darkmoon Tonk has the potential to finish the fight, but at that point it seems to be about 50/50 whether Nitun or the Tonk prevails. (It's possible I'm doing something horribly wrong of course, but so far as I can tell I'm following the guide's strategy exactly and my team just falls over before I can get Nitun down.)
The team I've had the best luck with so far is: Moon Moon (P/S) (2,2,2)Fossilized Hatchling (P/S) (1,1,1)Sandy Petrel (P/B) (2/1/2)
I doubt the breeds are ideal, that's just what I happen to have atm. Open with Moon Dance, if it'll let you survive a ravage to get off more bites, use Bite until Moon Moon is one turn away from death (don't let him get eaten by ravage; dying to rake is ok). Change to the Fossilized Hatchling, use ancient blessing if it'll help you survive an extra turn, otherwise just spam bite (usually I just spam bite). Sometimes Nitun dies to bite spam from the Fossilized Hatchling; if he doesn't, the Sandy Petrel usually can finish the job with one or two uses of Thrash. It's a bit less reliable than I'd like still (primarily in keeping Moon Moon from dying to an untimely Ravage), but it's worked well for me in general.
All that said, I also wanted to thank you for the tremendous amount of work you put into this; prior to getting a good team set up of my own, this guide helped me pick pets to level and was instrumental in me getting to a point where I could figure out alternatives to use in situations like this.... and Nitun is the only strategy here that I have had any difficulties with.
Darkwarrior42 wrote:Just wanted to post my experience regarding the listed strategy for Nitun.... burst team 2 is tagged "ideal" against him, implying that it should pretty much always work, but I rarely see it work.
An explode team is definitely what you want against Nitun since he never gets to kill anything with Ravage and never heals.
The other option to keep Nitun from healing, if you don't have an explode team, is to use an all Undead pet team. Because of the way that the Undead racial works, Ravage never gets credited as having killed an Undead pet and the heal will never trigger. Even though Nitun is a Critter, all of his attacks are Beast type so you won't take extra damage. Rikki's Bloodfang attack (3rd pet of the Celestial tamer Sully) works the same way which is why I like to use an undead as my second pet in that fight.
As a starting point, you could lead with a doom pet (Unborn Val'kyrs are always nice), follow that with a double damage debuff pet (Stitched Pup with Howl, Macabre Marionette with Dead Man's Party, Blighted Squirrel with Stampede) and just close out if needed with something that has a Beast type attack (your Fossilized Hatchling is a great choice).
Heck, any 3 undead pets with a beast attack should work. Bigglesworth is nice because when he Prowls his speed will drop below yours and you can Ice Barrier to block that attack (then just pass next turn to block the next one).
Here's another fun combo. Start with Blighted Squirrel and Woodchipper until you die; cast Stampede during your extra Undead round. Then go with Infested Bear Cub and Roar > Maul. You can kill him in five or six rounds with that. Nitun could disrupt that some by Raking your Maul but the third pet should be enough to finish Nitun off.
Another fun, though not as fast, combo I found was to start with Spirit Crab and Claw until your bonus round when you cast Rot to make Nitun Undead. Then you follow up with something like the Infected Fawn (Adrenaline Rush and Flurry) to finish him off.
Stencil wrote:Heck, any 3 undead pets with a beast attack should work.
Actually, I've realized you really only need 2 undead pets for Nitun. If he kills your third pet it doesn't really matter whether or not he healed up doing it; you've lost either way. So, you can probably put any hard hitter you want last.
|
{
"pile_set_name": "Pile-CC"
}
|
<?php
namespace LaravelFly\Providers;
class ConfigClearCommand extends \Illuminate\Foundation\Console\ConfigClearCommand
{
protected $signature = 'config:clear';
/**
* Execute the console command.
*
* @return void
*/
public function handle()
{
parent::handle();
@unlink($this->laravel->bootstrapPath('cache/laravelfly_ps_simple.php'));
@unlink($this->laravel->bootstrapPath('cache/laravelfly_ps_map.php'));
@unlink($this->laravel->bootstrapPath('cache/laravelfly_aliases.php'));
$this->info("LaravelFly ps configuration cache cleard!");
}
}
|
{
"pile_set_name": "Github"
}
|
How to bypass two-factor authentication facebook?
A comprehensive guide on how one can bypass Facebook two-factor authentication
Facebook, as you might know, is one of the popular social media platforms that allows users to connect with their loved ones located at far off places. Further, users can even utilize this platform to connect with the masses with common interests and likes. All in all, it is an incredible service to be socially active and reach out to the masses.
Indeed, the services offered by Facebook are the finest but there are chances that the account can be misused by a hacker. Fortunately, Facebook has devised some methods to prevent misuse of the account offering complete privacy and security. Well, there are multiple security measures introduced, still, two-factor authentication is considered as one of the safest ways to authenticate a user.
What is two-factor authentication?
Well, for the users who are not aware, the two-factor authentication is an additional security layer added to secure various online activities. In this process, the users are asked for a verification code once they sign-in to any of their online accounts.
In general, the verification code is provided to the user on their smartphone. Hence, adding an additional security layer to the online account at the time of login. Further, to get this security code, the user can use their phone number linked to their account. However, if the user encounters Facebook two-factor authentication without phone number, they can opt for third-party software like Google Authenticator or Authy.
Bypassing Facebook two-factor authentication
Undoubtedly, two-factor authentication is an incredible way to ensure the security of the Facebook account, but there are a few users who have reported an issue with the process because of lost access to all the authentication methods. Luckily, this issue is quite common and can be resolved in time, if the user opts for bypassing the two-factor authentication.
So, to help out the users who are looking for solutions on how to bypass two-factor authentication Facebook, here are the four effective methods that one can try out to bypass Facebook authentication.
1. Reset Facebook account password
One of the simplest ways to bypass Facebook account authentication is by simply resetting the account password. For the password reset process, the user is required to visit the Facebook login page and follow the steps provided:
Click on the Forgot password option and proceed.
Then, enter the email or phone number to find the account.
After that, the user can request a code on their registered email.
Further, the user is required to complete the verification process and reset the account password.
2. Using third-party to login Facebook
In case, if the user fails to bypass the two-factor authentication, using the above- mentioned method, they can opt for this alternative method. For this process, the user is required to login to their Facebook account using a third-party application like Google. So, for the users who have linked their Google account with Facebook, they can use this method to resolve Facebook two-factor authentication without phone number issues.
3. Use security codes assigned by Facebook
Besides, for the users who have opted for two-factor authentication for their Facebook account, they are offered with ten recovery codes that they can use for accessing their account if they have lost access to their linked phone number. By using this method, the user can easily bypass their account and regain access to their account in time.
4. Access Facebook account from recognized locations
And for the users who have logged in their Facebook account using another device or phone and still able to access their account from a recognized location can easily bypass the two-factor authentication. Once the user gains access to their Facebook account, they can make the required changes to their account and resolve Facebook 2 step verification not working issues.
Hence, these are the few methods that one can try out to bypass Facebook's two-factor authentication. And in case, if the user fails to gain access to their account even after trying out these methods, they can seek assistance from the help center and regain access to their account.
How can I recover my Facebook account without two-factor authentication?
Learn the procedure to recover a Facebook account without Two-factor authentication
Face book is the largest social media network in the world, a platform where a bunch of users is sharing the information. But certain times users unable to access their Facebook account, it may be possible when someone has gained the access to the Facebook account which potentially malicious intent. So stop worrying about anything or do not need to compromise your Facebook account. Here you will see all the simple and logical steps required to facebook account recovery without two-factor authentication.
Follow the steps to recover a Facebook account without two-factor authentication
With the help of the website or the service
In the first step effort, make sure that you have logged into the account.
Go to the official website of Facebook, and choose the simple way to recover the Facebook account.
In the next step, you will be redirected to the Facebook account recovery web page.
Moving ahead, Now tap on the recover account.
In the last step go for the security check-ups and get back to the service website.
Consider and adhere to the above information to get back to the account information Facebook account recovery without two-factor authentication. Even you can share the mail to the concerned department.
How to bypass login approvals on Facebook?
Failed to access your Facebook account because lost access to the code generator? Then, don’t need to fret out as there are alternative methods that one can try out to resolve this issue.
What is the need for a confirmation code for Facebook accounts?
For the users who are not aware, for accessing the confirmation code, it is required that the user has enabled two-factor authentication service that works as an additional security layer for the Facebook accounts.
However, many users have a query on how to bypass login approvals on Facebook as they have lost access to the phone linked to the account for generating code? So, to help the users with the query, here are some alternate methods they can try out.
Bypassing Facebook login approvals
For the users who still have access to the phone number linked to their account, they request to receive the confirmation code via text. Simply click on another way to authenticate and get a login approval code via text and regain access to the account.
Further, the user can use the saved recovery codes to get access to their Facebook account.
Try logging in to the Facebook account from an authorized device or location as there chances that the user might be still logged in and access their account and later make the required changes in the log in approvals settings.
How to remove two-factor authentication on Facebook without login?
Facebook is one of the popular social networking sites that has made it easier for people to stay in touch with loved ones located at far off places. Apart from this, Facebook even offers its users a platform to boost up their business and share their opinions globally.
And with such services offered, Facebook ensures the safety of accounts by offering users with some incredible security features including, two-factor authentication which adds an extra layer of security to the account.
However, there are a few users who have inquired regarding the process to disable two-factor authentication Facebook. So, to help users looking for this information, they can read out this article to gather information about the process.
Procedure to disable two-factor authentication
Initially, the user is required to visit the login page of Facebook.
Now, click on the Settings option under the Menu section.
Then, select the login and security option and click on the two-factor authentication option.
Further, scroll down to click on Edit and provide the Facebook password.
After that, the user needs to click on the two-factor authentication option.
And then, the user is required to confirm the pop-up and disable the feature.
Further, for confirmation, the user is required to click on the Submit button.
Thus, with the end of the process, the user can easily disable two-factor authentication facebook. For queries, one can contact the Facebook help center.
Post Comments ::
Comments ::
Buzzmeweb.com is an information provider website, where users can get all troubleshooting info about emails, printers, browsers and antivirus related technical issues. We using brand name, logo and more info only for information purpose. Buzzmeweb.com is not affiliated with any brand or company.
|
{
"pile_set_name": "Pile-CC"
}
|
Raymond Orpen
Raymond d’Audemar Orpen (31 August 1837 – 9 January 1930) was an Irish cleric in the 20th century.
He was a curate at Rathronan and then Adare before becoming the Incumbent of Tralee. He was Archdeacon of Ardfert until his ordination to the episcopate as Bishop of Limerick, Ardfert and Aghadoe in 1907. He retired in 1921.
References
Category:1837 births
Category:1930 deaths
Category:Alumni of Trinity College Dublin
Category:Irish Anglicans
Category:Archdeacons of Ardfert
Category:20th-century Anglican bishops
Category:Bishops of Limerick, Ardfert and Aghadoe
Category:Diocese of Limerick, Ardfert and Aghadoe
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Q:
Return XML Data up to next h2 only using XSL
I’m wondering how I can return only a portion of XML data up to the next h2 tag. So I have xml node called SUMMARY as shown below with some example text:
<SUMMARY>
<h2>heading One</h2><p>paragraph text under heading one</p>
<h2>Heading Two</h2><p>paragraph text under heading two</p>
<h2>Heading Three</h2><p>paragraph text under heading three</p>
</SUMMARY>
I currently am using this but it's not quite working
<xsl:choose>
<xsl:when test="contains(SUMMARY, ':')">
<xsl:value-of select="substring-before(SUMMARY, '.')"/>.
</xsl:when>
</xsl:choose>
Any help would be appreciated
A:
In XSLT 1.0, for retrieving content belonging to a specific milestone (in this case a heading element), I suggest one of the two following methods. Replace the <xsl:copy-of> part in these examples with whatever you want to do with the retrieved content.
1. Using keys:
<xsl:key name="content-by-heading" match="SUMMARY/p"
use="generate-id(preceding-sibling::*[self::h1|self::h2|self::h3|self::h4|self::h5|self::h6][1])"/>
<xsl:template match="h2">
<xsl:copy-of select="key('content-by-heading', generate-id())"/>
</xsl:template>
2. Iterating over the following siblings of the heading:
<xsl:template match="h2">
<xsl:apply-templates select="following-sibling::*[1]" mode="get-heading-content"/>
</xsl:template>
<xsl:template match="*" mode="get-heading-content">
<xsl:copy-of select="."/>
<xsl:apply-templates select="following-sibling::*[1]" mode="get-heading-content"/>
</xsl:template>
<!-- Stop iteration when we're at the next heading -->
<xsl:template match="h1|h2|h3|h4|h5|h6" mode="get-heading-content"/>
|
{
"pile_set_name": "StackExchange"
}
|
Correlation between miR-1207-5p expression with steroid-induced necrosis of femoral head and VEGF expression.
The aim of this study was to investigate the expression level of micro-ribonucleic acid-1207-5p (miR-1207-5p) in steroid-induced necrosis of femoral head (SNFH) and its correlation with SNFH. Meanwhile, we also aimed to analyze the relationship between miR-1207-5p expression and vascular endothelial growth factor (VEGF) in the femoral head. From May 2016 to December 2017, 60 patients aged (55.4±8.7) were selected in our hospital. All patients were diagnosed and confirmed as SNFH. Total RNA was extracted from the necrotic femoral head tissues and peripheral blood. Reverse Transcription-Polymerase Chain Reaction (RT-PCR) was used to detect the expression level of miR-1207-5p in tissues. At the same time, immunohistochemistry and Western blotting were adopted to detect VEGF expression in the bone tissue of patients with high or low expression of miR-1207-5p. 7 patients with femoral neck fracture aged (45.6±4.51) were enrolled in the control group. In the animal experiment, the rat SNFH model was established by intraperitoneal injection of lipopolysaccharide and methylprednisolone. Subsequently, the expression levels of miR-1207-5p and VEGF in necrotic femoral tissues were detected. Meanwhile, terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining was applied to detect cell apoptosis in bone lacunae of miR-1207-5p high expression group and miR-1207-5p low expression group, respectively. The expression level of miR-1207-5p in the necrotic bone tissue of the SNFH group was significantly higher than that of the control group. The expression level of miR-1207-5p was inversely proportional to Harris Hip score (p<0.05). A higher expression of miR-1207-5p indicated a lower expression level of VEGF (p<0.05). The animal experimental results revealed that miR-1207-5p expression in the necrotic femoral head tissue of SNFH group was significantly higher than that of the control group. Furthermore, the number of apoptotic cells in bone lacunae was remarkably higher in miR-1207-5p high expression group (p<0.05). MiR-1207-5p is significantly up-regulated in necrotic femoral head tissue and serum of SNFH patients. Meanwhile, its expression level is inversely proportional to Harris Hip score of patients. The possible underlying mechanism may be related to the inhibitory effect of miR-1207-5p on VEGF.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
When condo bylaws prohibit bike parking
The bylaws of older condominium associations often hinder the ability for communities to evolve as the needs of residents change. Some condo associations grapple with this issue as more residents start to ride bicycles.
Photo by Jeffrey Beall on Flickr.
Matt Johnson rents a condo in Greenbelt. His condo association's rules prohibit storing bikes on balconies. But there's nowhere else to store them outside; there are no bike racks outside anywhere.
They do have a "bike room" in the basement. It is completely unmarked; Matt just assumed it was an electrical closet for the first 3 years he lived there. After the building captain mentioned it to Matt, he went and looked inside.
There are no bike racks or any other elements to which it is possible to lock a bike. It's a big square room with cinderblock walls. Most of the room is full of junk from other residents, just stacked in there, no organization whatsoever. It looks like the only thing not stored in that room are bikes.
Instead, Matt keeps his bike in the dining nook during the warmer months. Whenever he and Ryan have guests over or during the winter, he moves it to their storage locker (also in the basement). It barely fits inside.
Even if they had a bike room, Matt probably wouldn't use it, because having to carry it up and down a bunch of stairs and going through 2 locked doors (for which each unit gets only 1 key) would inhibit daily use of his bike.
Veronica finds high hurdles to change her condo bylaws
When Veronica was president of her condo board in Fairfax Village, in DC's Ward 7, she looked into adding bike racks in their parking lot. However, the association's bylaws, written in 1974, prevent storing bikes outside:
Co-owner shall not place or cause to be placed in the public hallways, walkway, driveways, parking areas or other Common Elements any bicycles, furniture, packages or objects of any kind. The public hallways, walkways and driveways shall be used for no purpose other than for normal transit through them.
Once that was a no-go, she looked at the possibility of converting a basement space into a bike room. However, the same provision includes "other Common Elements." Another section of the bylaws has a half page definition of "Common Elements," which includes basements.
Amending the association's bylaws would require approval by two-thirds of owners at a meeting called specifically for that purpose. That's not an insurmountable task, but it would take a significant effort to get that percentage of owners at a meeting.
Given all of their other projects such as the community garden, bike parking fell to the bottom of the list. Fortunately, unlike with Matt's condo, most (if not all) of the residents that have bikes either own a garage unit or store them in a neighbor's garage unit.
DC's proposed new bike parking regulations might trump the bylaws, at least for some condos in the District. It's unclear if they would affect Veronica's condo, since the regulations only apply to buildings of 8 units or more, and Veronica's is 3 separate structures of 6 units, all connected together with plumbing. We emailed DDOT yesterday to find out, but hadn't heard a definitive answer by posting time.
Veronica O. Davis, PE, has experience in planning transportation, urban areas, civil infrastructure, and communities. She co-owns Nspiregreen, LLC, an environmental consulting company in DC. She is also the co-founder of Black Women Bike DC, which strives to increase the number of Black women and girls biking for fun, health, wellness, and transportation.
Matt Johnson has lived in the Washington area since 2007. He has a Master's in Planning from the University of Maryland and a BS in Public Policy from Georgia Tech. He lives in Greenbelt. He’s a member of the American Institute of Certified Planners. He is a contract employee of the Montgomery County Planning Department. His views are his own and do not represent the opinion of his employer.
Comments
Well, I think the condo board can re-think the intent of this particular provision and find an exception. The intent of this rule is to prevent random bike placement in common areas, not an organized, rule-based approach to providing parking for bikes.
Our building is putting in a bike rack. We don't have any interior room, and the parking spaces are all used, but we found an area on the property that was acceptable to residents. The only thing residents wanted to avoid was having the rack near the basement unit windows, which was completely reasonable.
All the condo board has to do is take an informal poll, put it up for discussion, and invite feedback. If the experience of our building tells anything, the support will be very strong for bike parking. If the board feels it really needs to bring this up an annual meeting, it's a headache but by no means an a major obstacle. In any event, it's a board is unlikely to go ahead and put in bike parking, and spend the money to do so, unless it has a clear go-ahead from residents.
River Place, a set of four co-ops in Rosslyn with over 1,400 apartments, has insufficient bike parking. In fact, River Place provides only covered spaces for about 100 bikes for the 1,400 apartments. This means bikes are parked outside and are often destroyed by the elements and vandals. It is remarkable that in 2014, we cannot secure enough bike parking...
Otions depend on whether the majority of residents support bike parking.
Along the lines of what kob wrote, if the majority supports bike parking, then the board should install bike racks at an appropriate location. Once installed, the bike racks are clearly not part of the parking lot or walkway so the rule against storing bikes would not apply. The rule is intended to limit the actions of individual residents, not the board.
If the majority is against bike parking, then it might take a local ordinance to void the convenant as contrary to public policy. That certainly has happened with no-pets covenants.
Temporary storage on balconies should be feasible with the right camouflage, though that might be more trouble than moving the bike.
Not to let detail get in the way here, but Matt RENTS his condo. He doesn't own it, nor does he have any right, perceived or otherwise to encourage or demand any concessions for anything from the association. If his landlord, the unit owner wants to make an issue of it, thats one thing, but I have a hard time getting wrapped around the axel when someone freely entered into a lease with someone that didn't have bike storage included, then decides after the fact that it is too bothersome to store the bike in the unit.
And the law, like the related law for commercial office properties is going to have so many (necessary) loopholes to make it worthless.
Perhaps you can add something to the residental DC building code for new construction, but I can't imagine most existing multifamily housing stock over~ 10 years old has the in building space to accomodate secured bike parking after the fact. I know the two older condo buildings I've lived in certainly didn't.
I prefer to store my bike in my unit anyway. I lived in a condo that had shared commercial parking and the bike storage area was behind a chain link fence with multiple racks. It only took 2 weeks for my first bike to get stolen and then only another 4 weeks for my second (replacement) bike to get taken. Ideally there would be an active (or even dummy) camera monitoring the area but that's just another expense that the association doesn't want to take on.
Ideally buildings would be constructed to have secure bike storage on the ground level. One apartment I lived in had done that. The bike storage room was adjunct to a general purpose storage room and assemble from both the main lobby and a rear entrance.
The room was well lit and clean and secured by fobs and cameras. It was also directly across from the buildings gym. The entire time I lived there no bike was ever stolen and I even left my pump beside my locked bike for convenience.
Then I moved into a condo where the "bike Room" was just space on the P1 level of the garage that the developer couldn't sell as parking. In this case it was under the entrance ramp. No fobs or cameras - just a chain link fence that wasn't locked. Predictably enough everybody's bikes were stolen in short order. The floor was littered with cut locks left by the thieves.
So they put a lock on the door to the "bike room". Now in order to get your bike you have to first go to the front desk and sign out a key. Take the elevator down to P1 to get your bike and then back up to return the key. If the desk clerk is busy you have to wait.
This still didn't stop the thieves. They just climbed over the fence as it didn't reach all the way to the ceiling. And the room is dirty. The ventilation fans blow constantly and the result is a thick muck that settles on the bikes within a week. So bad that nobody would ride such a bike without first throughly hosing it down.
The result is that everybody in my building is actually rides keeps their bike in their apartment. The only bikes now in the bike room are those in more or less permanent storage.
I grew up in NYC in a coop building that was built in the 1950s, and we had a bike storage room in the basement that was elevator accessible, and had wooden racks to store bikes in that were heavily used. This of course was back when it was assumed that families with kids would live in such a building, and every kid had a bike.
It was also, I guess, built with a full basement, and the underground parking garage was relatively small and had only I guess 1 space for every 3 units or so. The rest of the basement had the bike room, laundry room, party room, and storage areas.
I have to say I'm a little confused about the question posed here re by-laws and the actual law. I recall our condo association's lawyer telling us that our by-laws and rules cannot contradict DC and federal law. If DC has a law in place and all condos must abide by it; it does not matter what a particular condo's by-laws say. My condo used to ban anyone under 18. That was removed from the by-laws years ago when the Fair Housing Act was enacted. (And the change did not require a majority of ownership share approval like another by-law change would.)
Now, that doesn't mean the law can be easily accommodated by every condo. And that's what this column is really about I think. Now if only I could get our condo to remove all the abandoned bikes on our racks and in our bike room.
The reason this post got started was because I tweeted about the false dichotomy of the association, which is a pretty common (in my experience) dichotomy in residential developments.
Namely, if you store a bicycle where it is visible from the exterior of the unit, you can be fined because we don't want bikes to cause the value of the complex to decrease.
However, automobiles MUST be stored outside. They don't cause any problems whatsoever.
I personally wouldn't store my bike on the balcony anyway. I am perfectly happy to keep it in my dining nook. But I do think it's telling that the condo board thinks the mere sight of a well-maintained, clean bicycle is to be abhored, but there's no problem with parking a beat up, falling apart, non-working car (with Virginia plates) in the parking lot. (And yes, there is a car fitting that description in th parking lot right now. It belongs to my upstairs neighbor, who, as you might have gathered doesn't actually live in Virginia (since the unit above mine is also in Maryland)).
@Tina, yes they can be changed. But they usually require such a huge % of residents or ownership shares that it can be very difficult just getting enough people to vote. Our annual meetings for Board positions typically draws <30% of ownership shares even voting. We need >50% in favor to change our by-laws. Sounds like Veronica's might be even tougher.
It's tough to do bicycle storage well in an area of a building that's not designed for it. Security, access, liability, maintenance, 'common area' usage, expense are all real-life issues in a condo-- and you can't just brush them aside and announce that bike parking has to be set up immediately for a minority of condo unit owners. Note also that even under favorable circumstances, setting up things tend to take a long time in a condo. I'm not against bike parking-- I'm just saying it's a harder problem than you may realize.
River Place, a set of four co-ops in Rosslyn with over 1,400 apartments, has insufficient bike parking. In fact, River Place provides only covered spaces for about 100 bikes for the 1,400 apartments.
Where would you put additional indoor bike parking at River Place? The only place I can think of is the space in the South building that's rented out to commercial users but that's an important source of income for the co-op. For that matter, I don't even know where you put it outdoors unless you want to pave over green space.
I can't imagine most existing multifamily housing stock over~ 10 years old has the in building space to accomodate secured bike parking after the fact.
Totally agree although I wouldn't go so far to say "most". A lot of existing condos, for sure. It would be awesome to provide the additional amenity but short of acquiring additional land and building a purpose-built structure, I don't see how it's possible.
Not to let detail get in the way here, but Matt RENTS his condo. He doesn't own it, nor does he have any right, perceived or otherwise to encourage or demand any concessions for anything from the association.
Why not? The customer is king (or at least that's how any good business operates). I assure you that commercial tenants demand concessions from their landlords all the time.
I have a feeling that I live in the same condo complex as Matt Johnson. :-) I have been an owner there for nearly 15 years, but I have had issues with the condo president regarding how many flowerpots I put on my front patio, never mind bicycles. I live in a condo right by the main street, so yeah, appearances get scrutinized.
In my particular building, there is a bike rack in the storage room, but ever since I moved in (again, nearly 15 years ago) the rack has been filled to the brim with old, dusty bicycles that never seem to be used. There is no room for any other bicycles, and even if there were, we keep getting flyers from management saying that absolutely no personal belongings are supposed to be stored outside the individual storage bins, so I'm not sure I'd want to use the rack anyway.
I keep my bike inside my condo, but I wish I didn't have to. Then again, I grew up in the perfect little two-bedroom SFH with a basement and a garage, and I miss that terribly.
"In existing buildings, the [bicycle storage] spaces one for every three units must be provided within 30 days after a written request is received from one or more tenants or property owners." Seems like it should apply to condos.
But there's nothing in the law to require any particular features of the bike storage, so I wouldn't expect to see well-designed bike storage until demand exceeds some kind of tipping point.
Cages are better than nothing, but they MUST have a keyed deadbolt, or Jeffb's situation will be repeated. And I'm wondering how these these get access to the building? Can one simply walk down a ramp? Isn't that an overall security issue?
Moira beat me to it. The Bicycle Commuter and Parking Act of 2007, introduced by Tommy Wells, addressed this, and DDOT just finally issued new regulations that will require your condo to provide some bike parking, regardless of what the by-laws say.
Is it clear that the regulation applies to people who own a single unit in a condo building? Questionable whether they are a "residential building owner" but I don't see where that term is defined. Btw, the condo association doesn't own the building (the building is jointly owned by the owners as "tenants in common") although in a co-op, the co-op does own the building.
"Why not? The customer is king (or at least that's how any good business operates). I assure you that commercial tenants demand concessions from their landlords all the time"
Yes, of course it would be different if you owned the condo. You would have ownership into the association
And condo association is not Matt's landlord, and Matt isn't their customer. Matt's monthly rent goes to the unit owner, who as I clearly said before is clearly able to ask for these things if he wants to, but this isn't a rental company, and Matt isn't paying the association for anything, hence...no influence, nor reason for to expect any influence on the HOA.
Matt, you want the authority to demand such things, either buy yourself a condo and become a member of the association, or rent in a apartment building that has your preferred amenity.
I question whether the law or DDOT's proposed regulations would apply to condo or co-op associations, and would argue that they do not.
DC Code 50-1641.05(a) states that "A residential building owner shall provide secure bicycle parking spaces for all existing residential buildings with 8 or more units".
But a condo association does not own any units; each individual condo owner owns their unit. So unless a specific condo owner owns at least 8 units, it's not clear from a plain reading of the underlying statute that the DDOT regulations apply.
The underlying statute was not well drafted (it left off any mechanism for enforcement until that was added a few years later through an amendment) and it's not clear the statute was intended to apply to condo associations, rather than rental apartment buildings. If you read the Committee Report on the Council's website, it's not clear what the Council's intent was on this issue.
Instead of setting up a potential legal challenge, DDOT should clarify that these proposed regulations are applicable only to rental apartment buildings that are owned by the same person/entity and have at least 8 residential units.
Additionally, DC Code 50-1641.07(b) requires these proposed regulations to be submitted to Council for review and approval. But DDOT's proposed regulations don't make any mention of that. Was that an oversight?
""In existing buildings, the [bicycle storage] spaces one for every three units must be provided within 30 days after a written request is received from one or more tenants or property owners."
This wouldn't help Matt in any case. He isn't a property owner, and he is not a tenant of the association. Matt's landlord could ask for it, but not Matt.
And as it relates, this law will have so many waivers and loop holes (as does the commercial office regulation for bike rooms and showers)to make it next to useless. If I live in a 10 unit building that simply doesn't have extraneous empty rooms built into the first floor (I mean really, what average condo or apartment building does?), what is the HOA supposed to do? Spend tens of thousand of dollars, designing, permitting and building an add on room to the building for bike storage? Never happen.
I imagine that a lot of buildings will have to seek waivers to this law. Our building has looked at the issue, and it's impossible to provide interior parking. There is no space.
There is always the possibility of installing outdoor container-type units, similar to what is at Metro stations, but these units are expensive. We are putting in an outdoor rack, but I don't think it meets the law. Even if an acceptable location could be found for bicycle storage units, the condo board (which i once served on) will be hard-pressed to spend the money to buy them.
Does the law require provisions for bicycle parking to be free? Because if storage containers were forced upon the building, I'm certain the board would insist, and rightly so, on charging enough rent to amortize purchase cost and maintenance.
The condo rule referenced in the story is about keeping bicycles out of the hallways, where they are potential hazard, and other common areas. They weren't written to prevent boards from creating common storage areas. The interpretation that they constitute a blanket ban on bicycle storage is just wrong.
I think Wells would have been better off requiring new building to provide for bicycle parking; he probably knows that most older buildings are space constrained, and every inch of most buildings is to put use already. The only thing this new does is impose a BS requirement on condo boards.
I meant Matt has a right to demand concessions/bike parking from his landlord -- the unit owner -- not the association. There's nothing in the post or comments that indicates Matt was making demands directly to the HOA. Actually, I don't see where Matt is making demands of anyone. All I read was him pointing out the absurdity of the situation to GGW readers.
If a condo has no place to put bike parking, than banning keeping bikes on a balcony seems quite unreasonable, at least to me. Of course a condo association has the right to be unreasonable as long as they comply with the law (thus this will be different in DC than in Md or Va)
The recourse (aside from appealing to elected officials to change the law) is to ask the condo owner to get involved. In a bikeable place being unable to keep one's bike on a balcony (in a building without bike parking) may well diminish the rentability of that unit. Unfortunately many absentee owners of condos (and houses in HOAs) are not active in the condo assoc or HOA.
If the building/HOA is small and intimate enough, one might be able to appeal to neighborliness and common sense. In a larger one it might be feasible to make common cause with owners who have bikes.
"This wouldn't help Matt in any case. He isn't a property owner, and he is not a tenant of the association. Matt's landlord could ask for it, but not Matt."
If Matt can find any owner in the assoc, whether a personal friend, a fellow cyclist, or whomever, willing to request it, the law would come into effect, IIUC. Even if his own landlord is not willing to.
"That's not an insurmountable task, but it would take a significant effort to get that percentage of owners at a meeting. "
wait, its 18 units, and you need 2/3 - so 12? Veronica plus 11 others? Is there anything in the bylaws preventing a member from bringing food to a meeting? Because I think that for a modest expenditure (or effort in the kitchen) a turn out of eleven neighbors should be doable.
""In existing buildings, the [bicycle storage] spaces one for every three units must be provided within 30 days after a written request is received from one or more tenants or property owners."
This wouldn't help Matt in any case. He isn't a property owner, and he is not a tenant of the association. Matt's landlord could ask for it, but not Matt.
This opens up a whole can of worms. Can the written request come from a tenant in a condo building or does the request need to come from tenants in apartments and owners in condos? Who would the owner of a condo unit submit their request to? The HOA doesn't own anything and is a "building owner". Does the property owner making the request need to be a resident of the building?
@Lurker
Agree the intention is not clear but the fact the regulations stipulates a request can come from one or more property owners (plural), makes it seem like they wanted it to apply to condos. That's the main situation where there would be multiple property owners.
Also, what's crazy is that the regulation would apply differently to co-ops because a co-op is a corporation that owns the entire building and the shareholders (similar to unit holders) own shares in the corporation. So, in that case the corporation is clearly the residential building owner. However, the shareholders are neither tenants nor property owners so they could not submit a request according to the regulation.
Seems like this regulation is poorly though out and written. If it's meant to apply to condos, the main effect will be to cause them to go through a lot of paperwork to apply for a waiver because as others have noted, most buildings don't have extra indoor space lying around that can be used for bike parking.
In a condo, each individual owner owns his or her unit and the common areas are owned jointly by all the unit owners. So by my understanding the residential building "owner" refers to all the unit owners as a whole. (Note that the law does not specify that a residential building owner must be a single individual, but rather merely an "owner".)
The Code probably doesn't require a condo association to carry out the law because the Code does not require that every condo building form an association. But most condos do have associations because that's the most practical way for a building's "owner" to manage or operate the property.
"I find it pretty funny that folks like AWINTC will wax poetic about the waste and injustice of forcing a developer/owner to eat the cost of vehicular parking. "
No where here have I addressed the desirability of a bike parking minimum vs allowing the market to decide.
I do think that by condo/HOA bylaws banning bike parking, are particularly unreasonable, as rescinding them costs the building nothing. Imagine an HOA where everyone had a paved pad next to their house, but storing a car there was banned because the HOA thought that cars did not look attactive. I would think that complaining about such an HOA would be in a different category from calling for parking minimums.
Note that Fairfax has no such minimum (AFAIK) and I am not calling for one. I will comment on DC's bike parking minimum only if a thread is started where that policy question becomes relevant.
BTW if you are truely concerned with congestion on the dulles toll road, and in Fairfax generally, and with solutions that are cheaper than more transit projects, you should certainly be encouraging more biking.
But man o man, that "injustice" goes right out the window when its forcing developers/owners to eat the cost (tens of thosands)in building extra rooms, with cameras and security for your bike.
I haven't read it but I'm pretty sure the DC law has a grandfather clause meaning that for existing buildings its up to the owners to decided to add a bike room or not. Or, they could allow bikes to be stored on the balcony or somewhere else.
Also, most of us on this site (including AWITC because he/she has said so before) that don't like car parking minimums aren't doing it because we have a general anti-subsidy position. We don't like the policy because we think its bad or ineffective.
You've been very vocal about how it is ridiculous to require developers to incorporate parking because the costs filter down to the tenant/ buyer. This is exactly the same.
How many people do, or would bike from Leesburg to their job in Tysons corner or DC every day.
HOA's disallow thing on their common area elements for numerous reasons, asthetics and liabilty being the biggest two, and I (as a one time condo owner) certainly don't fault them for it. The rules are there prior to move in, so if someone doesn't like it, one would wonder why they agreed to live there in the first place. Kinda like buying ahouse next to a fire station, then constantly complaining about the noise.
Mostly though, I think the solution is not via legal minimums, but by organizing within communities. The solutions Matt requires (being allowed to keep his bike on his balcony when he chooses) and that veronica wants (bike racks in surface parking lot) are hardly very costly, and are quite feasible. Its hard to beleive that reasonable human beings would object - the problem is with certain institutional factors - in Veronica's case the difficulty in altering by laws - in Matts, the lack of influence of a renter on a condo assoc. I think both can be overcome without using a law requiring bike parking - and if the law does help (in veronicas case) it would do so in a way requiring no financial hardship on the assoc.
kob - While I agree of the intent, our by-laws have caused enough legal battles that we decided to leave it alone. For example, when I was president the board wanted to charge for parking in the lot. There was a legal battle for 2 years as to whether the board had the standing to charge for parking and if it was a violation of the by-laws. We argued that we weren't charging for a parking space, but an administrative fee for the parking permit. After two year legal battle over interpretation, the board won. It was expensive and time consuming.
AWalkerInTheCity - My condo association is 85 residential units + 36 garages (garages are considered a ownership unit). Each unit owner has a percentage ownership based on the size of their unit. For example, I am a 1.4% owner because I own a residential unit and a garage unit.
"You've been very vocal about how it is ridiculous to require developers to incorporate parking because the costs filter down to the tenant/ buyer."
I have said they are undesirable for that reason - and that they fail a benefit cost test.
" This is exactly the same."
Maybe - I am not sure if they are the same from a cost benefit POV. Many jurisdictions want to encourage more biking - even those with parking minimums generally are not trying to encourage more biking. There may be exceptions though.
"How many people do, or would bike from Leesburg to their job in Tysons corner or DC every day."
How many would bike from Reston to Tysons? Vienna to Tysons? Vienna to North Arlington? Ashburn to Reston? Ashburn to Vienna?
"HOA's disallow thing on their common area elements for numerous reasons, asthetics and liabilty being the biggest two, and I (as a one time condo owner) certainly don't fault them for it."
I think seeing a bike on a balcony makes a building seem MORE desirable ;) I certainly can't see why a bike is less desirable than a grill (does Matts building allow those, or the many other things routinely found on balconies?)
"The rules are there prior to move in, so if someone doesn't like it, one would wonder why they agreed to live there in the first place. Kinda like buying ahouse next to a fire station, then constantly complaining about the noise."
If there is a cost free way to reduce fire station noise, and its rejected out of petulance, that would seem unreasonable to me. Caveat Emptor should not, IMO, preempt logic and reason. And certainly not neighborliness.
BTW - does Matts condo have parking? If matt were to keep his bike on a bike rack on a car, would that be allowed?
veronica - okay, the article was confusing then. Im not sure 85 would be hard to organize (so youd be okay with 60 attending) if you have a bunch of cyclists who would help do the organizing. But it sounds like you don't?
So by my understanding the residential building "owner" refers to all the unit owners as a whole. (Note that the law does not specify that a residential building owner must be a single individual, but rather merely an "owner".)
That's a reasonable interpretation. In that case, who would be responsible for the fines for non-compliance? All the owners in proportion to their ownership percentage including the owner who made the request? If the fine went unpaid, would the city place a lien on everyone's units and whoever's unit got sold next would end up paying the entire fine out of their sales price?
I have suggested parking minimums generally fail cost - benefit tests in locations with good transit. I have NOT said all parking minimums are undesirable. In places where high density uses are located far from transit, parking minimums are much more like to make sense. Personally I think locating such uses far from transit is a poor idea, but it does happen.
Awalker in the City - It's garden style condos/townhouse complex. We have 3 building clusters (21 buildings total). My building is physically attached to two other buildings (i.e. shared plumbing and all) and there are a total of 7 buildings my cluster. Hope that clears it up
Not yet. Also, I don't think you can be grandfathered in under these regs. Originally you could, but a law was passed to change that a few years ago. I could look it all up, but it's really confusing between the law, the zoning regs, the Sustainable DC plan etc... It gets kind of jumbled.
And note I have not come out one way or the other WRT to parking maximums. To the extent that the argument against parking mins is made purely on free market/econ 101 principles, the same arguments apply to parking maximums, of course. But reducing or eliminating parking mins is often suggested in places where its public policy to encourage mode shift to non-auto modes. IN place with high congestion, a parking maximum (or a bike parking min) can serve a purpose that is also served by relaxing parking minimums. In each case the impact on rents and the external impacts need to be weighed.
However, regulations cannot go beyond what their enabling statute authorizes.
In this case, the enabling statute makes clear it is applicable to "a residential building owner" with 8 or more units. Note the singular description of owner.
Therefore, DDOT cannot issue regulations that go beyond what is authorized by the underlying statute - which is what I believe a requirement on condo and coop associations would be.
I would argue that if the statute was intended to apply to condos and coops, it would have used a phrase like "all residential buildings" and not focused on the ownership of the building. But that's not how the statute is written.
Because there is no grandfathering for this law or these proposed regulations, DDOT is likely going to get flooded with requests for exemptions. And, under its proposed regulations, DDOT provides no due process relief if it rejects an owner's request for an exemption or variance. Can an owner appeal? Where? How? None of those questions are answered.
Finally, assuming that these proposed regulations were deemed by DDOT to apply to condo associations, who would receive fines for noncompliance? The condo association? The condo board? The individual condo owner who can't act outside what's authorized by the condo association and its bylaws?
Again, it seems that the easiest way for DDOT to avoid causing a significant mess for many condo and coop associations is to simply limit these regulations to rental apartment buildings where the ownership structure is crystal clear and where there is no statutory confusion about the law's applicability.
Lurker, I can assure you, these regulations were always intended to apply to condos. Condos are residential buildings. Condos have owners. Therefore "a residential building owner" applies to condos. Who owns the condo's common areas if not the condo unit owners? And who makes decisions about those areas if not the condo association? Whoever is responsible for addressing other regulations like ADA and fire codes in condos is responsible for addressing these regs. I think you're parsing the language to such a fine point that you're getting ridiculous. Perhaps the courts will side with you, but DDOT believes these regs apply to condos and until they are told otherwise, will enforce them as such.
and, under its proposed regulations, DDOT provides no due process relief if it rejects an owner's request for an exemption or variance. Can an owner appeal? Where? How? None of those questions are answered.
The same way other exemptions, variances and appeals are dealt with. I assume that is zoning.
In this case, the enabling statute makes clear it is applicable to "a residential building owner" with 8 or more units. Note the singular description of owner.
So you believe that the statute only applies to buildings with one individual person as an owner? I guess that releases rental buildings (which are usually owned by corporations) from responsibility to adhere to the law as well. Oops.
Or maybe you're simply reading the statute too narrowly. What prohibits the term "owner" from describing a plurality of individuals that share ownership of a property, just as they might share ownership of a car, a bank account, etc?
I'm not parsing the language at all; I'm simply reading the law as it is written in the DC Code, without the assurances by a online commentator of what an agency "really" meant when it promulgated regulations.
And the examples of fire codes and ADA compliance (which both have grandfathering provisions) simply buttresses my point: They are applicable to all residential buildings, regardless of the ownership of the building.
Rather than have a never-ending online comments merry-go-round of what the definition of "a residential building owner" is, it's far simpler and more productive for DDOT to publicly state whether they intend it to apply to condo and coop associations.
And, since the underlying law requires Council review and approval of the regulations, it would ultimately be up to Council to decide whether they agree with DDOT's interpretation.
@ David C - No, due process appeals arising under these proposed regulations aren't like zoning at all; there is no authorization for either the BZA or the Zoning Commission to hear any such appeal because it wouldn't arise under either the Construction Codes or the Zoning Regulations.
A civil infraction can be appealed to the Office of Administrative Hearings (although it's not clear to whom a fine would be issued in a condo or coop association). But that's for a civil infraction, not for a rejection of an exemption/variance request.
DDOT needs to spend some more time working on these proposed regulations; they are not fully thought through, as demonstrated by the conundrum faced by Ms. Davis' condo association. They need to get more public input on its practical impacts.
Whoever is responsible for addressing other regulations like ADA and fire codes in condos is responsible for addressing these regs.
Condo associations are responsible for addressing ADA, etc. The bike regulation applies to building owners and the association is not the owner. Building owner is not defined.
If you want the regulation to apply to HOA's, you need a line such as:
"Under the Regulations, the Condominium Association, Cooperative or Mutual Housing Corporation, is for the most part considered to have the responsibilities of the owner as described in the Act, and is responsible for the registration of each building and for the correction of all cited violations."
DDOT believes these regs apply to condos and until they are told otherwise, will enforce them as such.
DDOT is going to be the agency enforcing this? What part of DDOT will be visiting all the residential buildings in the city to do inspections?
@Scoot - A rental building has a single owner; whether it is an individual or an LLC or an incorporated company, it is still a single owner. That single owner is easily found in the records of the Office of Tax and Revenue since it's the entity that pays the property taxes.
I am reading the statute as written; it refers to "a residential building owner." There is no "a" single residential building owner that owns more than 8 units in a condo association.
" What part of DDOT will be visiting all the residential buildings in the city to do inspections?"
The Fairfax County office of code compliance has recently stated that it only enforces maximum occupancy rules when someone complains. I imagine this will be similar (and say 10 unit condos where NO ONE wants bike parking will be effectively exempt) Of course you still need someone to enforce when there IS a complaint - but inspection of ALL buildings is hardly required.
A civil infraction can be appealed to the Office of Administrative Hearings (although it's not clear to whom a fine would be issued in a condo or coop association). But that's for a civil infraction, not for a rejection of an exemption/variance request.
Ok, well we could just read the code:
"Where it can be demonstrated that providing sufficient bicycle parking spaces required under § 1214.2 or § 1214.4 is not physically practical, that undue economic hardship would result from strict compliance with the regulation, or that the nature of the building use is such that bicycle parking spaces would not be used, the District Department of Transportation Bicycle Program Office may grant, upon written application of the owner of the building, an appropriate exemption or reduced level of compliance. In such cases, the Bicycle Program Office shall issue to the building owner a written certificate documenting the exemption or reduced level of compliance"
Fine, a building has to provide parking. However, a building is not a legal entity that is subject to laws anymore than a dog can be taken to court. A corporation, person, or other similar entity is what is subject to laws (in the case of dogs, it's pretty clearly stated that the dog's owner is responsible for say a dog bite). So a building has to provide bike parking? Who or what is responsible for ensuring the building provides the parking? The "building owner". What constitutes a building owner? The reg doesn't say.
" What part of DDOT will be visiting all the residential buildings in the city to do inspections?"
The Fairfax County office of code compliance
Yeah, I hope it's DC's office of code compliance and not DDOT that will be doing enforcement.
it's far simpler and more productive for DDOT to publicly state whether they intend it to apply to condo and coop associations.
Well, I doubt anyone from DDOT is reading these comments, so maybe the solution would be for you to reach out directly to someone with the authority to definitely answer your question. That way your confusion may be revolved once and for all.
I won't lose much sleep either way as my condo already provides bike storage.
FWIW, it's difficult retrofitting bike infrastructure into the sidewalk-roadway world too, which wasn't originally built to accommodate it.
The thing is to ensure the best possible parking when buildings are first constructed, because once a building moves to the delivered phase, getting a condo board to agree to (re)new parking and spending takes a long time. But some condo buildings are doing it.
My experience in trying to get developers to do the best possible parking in new buildings rather than meeting the minimum requirements is very very very very hard, even if the developers are seen as otherwise "enlightened."
There are other options for retrofitting if there is space on site, to include underground options as well as above-ground options.
But I agree that ideally space should be provided at the ground level. I got into an argument with someone from a real estate consulting firm at a conference when she talked about a project in Denver that she thought was really cool because it included a wall rack for a bike in every underground parking space. I said it was great for storing bikes but not for using bikes, depending on difficult it might be to get to the parking from outside.
LEED-ND is providing somewhat better guidelines than most cities in that it includes guidelines about convenient access.
I also think that in dense neighborhoods cities can do more. In the Netherlands they have higher quality bike corrals with protection in street "parking" spaces. I can see that in the city. As well, again, as shared underground parking.
While I appreciate all the interest in the DC regulations, those regulations will never cover communities that are not part of the District.
To be honest, there are lots of non-bicycle-related improvements I would like to see to bring my 50-year-old Greenbelt condo complex into the 21st century. A partial list: solar panels on south-facing roofs, a clothesline area for residents to hang laundry out on sunny days, a few parking spaces with charging stations for electric cars, and maybe even a Zipcar, so that people who can't or don't drive are not stranded at home on Sundays, when Greenbelt has no bus service. Oh, yeah, better insulation in the units too.
However, I'm realistic. The way our community is laid out, there's no good place to put a clothesline area without having it seen from someone's front door. And I haven't run the numbers on solar energy.
We aren't supposed to have grills on our balconies and patios either, but I believe the City of Greenbelt forbids open fires within 25 feet of a building, so there's not much the HOA can do about that.
Maybe someday I will run for the association board, but I need to find more gainful employment first. I can't be worrying over the condo association if I'm stressed out over my bills and my fluctuating income.
The bike regulation applies to building owners and the association is not the owner.
My wife, who has defended Condo Associations many times, says that condo owners own the building with each owning a "divided share" of it. The building has bylaws to govern how the powers and obligations of that ownership are exectuted - usually through a condo board. They act as the owner and are considered as such.
So usually it is the condo board, for all legal purposes, the owner. But check your buildings by-laws.
BTW, if anyone really thinks that their 20 minute perusal of the regulations - which once adopted become law - are missing some critical point (despite all the public reviews and work done by city lawyers, DDOT and council staff as well as the office of the Mayor) we are still in the comment period. The link I gave above has a link for comments. They really want to hear all the ways that they don't understand about condo ownership.
And if you look at section 1216 it mentions the fines and refers you to other sections that deal with how they are given out and adjudicated, which may include appeals.
"I have had issues with the condo president regarding how many flowerpots I put on my front patio, never mind bicycles." Yeah, that's not unusual for condos.
If you've already tried politeness, if you got any legal basis you can cite meaning you can keep the bike on your patio, wave it at them. Otherwise, they probably will be mean to you. sorry.
@ David C - I'm sorry, but it is not legally accurate to say that a condo board is the owner of the condo association. The DC Condominium Law is quite clear on how ownership works: there is no *single* owner of a condo building.
But you are correct that people should submit comments to the proposed regulations. It would not be the first time that DDOT - or any other government agency - has promulgated regulations without understanding their full impact or unintended consequences. It's a shame, though, that DDOT doesn't list all received public comments to its proposed regulations on its website. That would be a very useful public service - and also save their FOIA officer a large amount of time.
Lurker, I did not say that the condo board is the owner. I said that they act with the power of an owner (in cases where the by-laws say so). Why don't you show me the parts of the DC Condominium Law that is so clear that I am wrong.
@ Gabriel, @ Falls ChurchRiver Place, a set of four co-ops in Rosslyn with over 1,400 apartments, has insufficient bike parking. In fact, River Place provides only covered spaces for about 100 bikes for the 1,400 apartments.
At River Place North, we recognized the need for more bike parking. With nowhere to expand as Falls Church notes, we went vertical in the existing bike room.
We consulted with bike room users, BikeArlington, and the Police on better room usage and security. Then we looked at potential solutions and got bids.
The room now accommodates almost double the bikes. The configuration has restored order and added maneuverability. It uses 'double decker' racks where some bikes are elevated. The elevated racks are more difficult to use for some and preferred by owners of nicer (and lighter) bikes who didn't like the chaos of the previous configuration.
We moved the old schoolyard style bike rack outside for visitors and quick stops. This required Owners Assoc. approval since it manages outdoor space and chains illegally parked bikes. While the bike room requires registration, the outside rack is open for all to use.
For security, bike room users currently need to get the bike room key from the office or front desk. Cameras cover both the bike room and the outdoor rack. We're moving to a key fob system next month so we'll try fob entry on the bike room - the system records each entry and we can sync entry time stamps with camera footage. We will add access restrictions only if necessary for security. If none of this works for residents, bikes can be transported on any elevator. We ask that bikes come in/out through the loading dock or one of the side entrances instead of through the lobby.
Gabriel - All RP buildings are interested in resident input, whether you are an owner or renter (both are welcome at board meetings). If you decide to request more storage from the management or board of your bldg, you/they are welcome to see what we've done at RPN if helpful.
"Now if only I could get our condo to remove all the abandoned bikes on our racks and in our bike room.
by RDHD on Mar 26, 2014 12:21 pm"
I think this comments says it all. Abandoned bikes! My condo bldg. has a dedicated area in the garage and every year we have to go through the process of people identifying their bikes. If they aren't identified then they are removed. Bikes attached to lightposts and then abandoned. I sympathize with the problem of bikers, but I don't want to see someone's bike on their balcony and I don't like to see 2-3 stripped bikes attached to lightposts.
I sometimes prefer to keep nicer bikes inside my apartment, mostly because the process of getting in/out of the bike room takes too long. (Bikes don't get garage door openers, but of course cars do.) Even though I double-checked the house rules beforehand to make sure it was OK, I've gotten shade from some of my neighbors for deigning to bring my bike onto the elevator. I'm always super courteous about it (I've never so much as brushed by anyone), and I keep my bike a lot cleaner than some neighbors keep their dogs...
If the reason why the rules can't be changed is an inability to get the correct number of people at the meeting....
...change the bylaws or constitution to reduce the number of people required at the meeting. It doesn't really matter whether you do this with the right number of people present at the meeting. The people present and voting simply vote to waive the quorum rules and to waive the attendance rules, and then start passing changes.
There's only one way for people to object to this: *enough people have to show up at the next meeting*. If that happens, it solves your problem.
Add a Comment
Name: (will be displayed on the comments page)
Email: (must be your real address, but will be kept private)
URL: (optional, will be displayed)
Your comment:
By submitting a comment, you agree to abide by our comment policy. Notify me of followup comments via email. (You can also subscribe without commenting.) Save my name and email address on this computer so I don't have to enter it next time, and so I don't have to answer the anti-spam map challenge question in the future.
|
{
"pile_set_name": "Pile-CC"
}
|
Fluoroscopy radiation safety for spine interventional pain procedures in university teaching hospitals.
Fluoroscopic guidance is frequently utilized in performing various types of interventional techniques. The major purpose of fluoroscopy is accurate needle placement to ensure target specificity and accurate delivery of the injected drug. However, radiation exposure may be associated with risks to physician, patient, and personnel. Multiple studies have evaluated the risk of radiation exposure and techniques to reduce the risk in private practice settings. However, the literature is scant in evaluating the risk of radiation exposure in teaching hospitals in university settings. To evaluate safety and duration of radiation exposure for fluoroscopy guided interventional pain procedures in university pain clinics. Retrospective, case study. The data was reviewed from the fluoroscopy machines from March 2004 to April 2004 at two university pain clinics. Mean fluoroscopy time (FT), mean radiation dose per procedure, and utilization of pulsed fluoroscopy were analyzed. Data of a total of 165 cases of spine injection procedures were collected. The mean fluoroscopy time for lumbar epidural steroid injection, facet joint block, sympathetic nerve block, sacroiliac joint injection, and discography were 46.6 +/- 4.2; 81.5 +/- 12.8; 64.4 +/- 11; 50.6 +/- 41.9 and 146.8 + 25.1 seconds respectively. There were significant differences in fluoroscopy exposure times and radiation dosage for epidural steroid injection among different teaching physicians. Pulsed fluoroscopy was used in less than 10% of cases. The results of this study show that the fluoroscopy exposure time for various interventional procedures performed in the university settings are significantly higher than the radiation exposure periods in private practice settings. This study also showed significant differences among physicians in the same university setting.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Q:
Change the status automatically after the add of a revision
I have 2 forms, the first is the form motorbikes with 3 fields (matriculation, number_motorbike, status).
Then, we have the form revisions with 4 fields (date_revision_start, date_revision_end, garage, fk_motorbike)
Here in the form motorbikes the status of the number motorbike must to be automatically unavailable.
I must to work on which controller to automate the availability status of the motorbike?
Edit: 09 / 07 / 2018
Controller Motorbike
public function store(Request $request)
{
$bikeIdsDown = Revision::where('date_revision_start', "<", Carbon::now())->where('date_revision_end', ">", Carbon::now())->pluck('id')->toArray();
return view('motorbikes.index', compact('motorbikes', 'bikeIdsDown'));
}
Index.blade.php
@foreach($motorbikes as $bike)
<tr>
<td>{{ $bike->martriculation }}</td>
<td>{{ $bike->number_motorbike }}</td>
<td>
@if(in_array($bike->id, $bikeIdsDown))
UNAVAILABLE
@else
Available
@endif</td>
<td>
<form action="{{ route('motorbikes.destroy', $motorbike->id) }}" method="post">
{{ csrf_field() }}
{{ method_field('DELETE') }}
<a href="" class="btn btn-primary">Details</a>
<a href="{{ url('motorbike/' .$motorbike->id. '/edit') }}" class="btn btn-warning">Editer</a>
@csrf
@method('DELETE')
<button type="submit" class="btn btn-danger">Supprimer</button>
</form>
</td>
</tr>
@endforeach
A:
To show the list of motorbikes that are unavailable, you would need to work on your Motorbike Controller. Specifically, the edit() method if you are making changes to the status manually, the create() method if you are making a new motorbike, or most likely, the show() method to just display that table you have above where it shows if they are available or not.
Now... to get the part where it displays if the bike is unavailable... I assume it is unavailable when it is down for repair (revision). Since I don't know what your relations are, we can figure this out using a simple pull from all revisions to see which bikes should be down currently. So, in one of those methods I noted above (or all of them), first let's see which bikes are down for revision, during the current time:
$bikeIdsDown = Revision::where('date_revision_start', "<" Carbon::now())->where('date_revision_end', ">", Carbon::now())->pluck('id')->toArray();
Not the most efficient, but simplest to follow hopefully.
Now that we have these unavailable motorbike ids, if we compact them and send them through to the blade page, we can determine whether to write 'available or 'unavailable' in the motorbike field. So, as you are looping through the motorbikes on the blade page, maybe something like this:
@foreach($motorbikes as $bike)
{{$bike->martriculation }} // <-- these might all be in <td> or something
{{$bike->number}}
@if(in_array($bike->id, $bikeIdsDown))
UNAVAILABLE
@else
Available
@endif
@endforeach
Hopefully this will give you the idea.
|
{
"pile_set_name": "StackExchange"
}
|
Fun with statistics 2: The slightly-less-mythical median liblog
That post, which like this one is part of a non-series leading up to “But Still They Blog,” gave the metrics for the “average liblog” from the population of 521 fairly visible liblogs in the study. In that case, not only was there no such liblog, it would be impossible for there to be such a liblog, as some of the metrics conflicted with others.
So let’s look at the “median liblog,” taking the median instead of the average for those same metrics.
General characteristics
The blog began in November 2005, it has a Google Page Rank of 5 (or had such a rank in the summer of 2009), and when checked on September 30, 2009, the most recent blog was more than one week but less than two weeks old.
2007
The blog had 19 posts in March-May 2007, totaling 4,028 words in posts averaging 206 words each. There were nine comments, or 0.5 comments per post.
2008
The blog had 17 posts in March-May 2008, totaling 4,098 words in posts averaging 230 words each. There were 11 comments, or 0.7 comments per post.
2009
The blog had nine posts in March-May 2009, totaling 1,990 words in posts averaging 202 words each. There were five comments, or 0.4 comments per post.
Changes
There were 23% fewer posts in 2008 than in 2007, 31% fewer in 2009 than in 2008, and 34% fewer in 2009 than in 2007.
While the blog as a whole was 6% shorter in 2008 than in 2007, 31% shorter in 2009 than in 2008, and 35% shorter in 2009 than in 2007, posts were within a percentage point of being the same length in each year.
While there were 31% fewer comments in 2009 than in 2008 and 34% fewer in 2009 than in 2007, comments per post were within one percent of being the same each yer.
I think the individual-year metrics are internally consistent or close to it–but some of the changes aren’t internally consistent. There’s a simple reason for that, having to do with default values for empty cases.
Silliness
This is all silliness, of course–and the project itself does not include any of this. If I did describe a median blog, it would be done using trimmed medians (omitting empty cases)–but I doubt that I will.
One question is more interesting, and I don’t have an answer yet (and won’t for a while): When you divide metrics into quintiles (e.g., the 20% most prolific blogs, the next 20%, the middling 20%, etc.), will there be blogs that are “middling” on all measures–that fall into the third quintile for all metrics? We shall see…
This entry was posted on Friday, October 9th, 2009 at 2:42 pm and is filed under Liblogs. You can follow any responses to this entry through the RSS 2.0 feed.
Both comments and pings are currently closed.
*All Cites & Insights PDF ebooks are explicitly site-licensed for
mounting on a library's server and providing to authenticated users. That
includes The Gold OA Landscape 2011-2014, A Library Is..., Beyond the
Damage and any others.
|
{
"pile_set_name": "Pile-CC"
}
|
// SPDX-License-Identifier: GPL-2.0-or-later OR MIT
/dts-v1/;
#include <dt-bindings/gpio/gpio.h>
#include <dt-bindings/input/input.h>
#include "ar9331.dtsi"
/ {
model = "Pisen TS-D084";
compatible = "pisen,ts-d084", "qca,ar9331";
aliases {
serial0 = &uart;
led-boot = &led_system;
led-failsafe = &led_system;
led-running = &led_system;
led-upgrade = &led_system;
label-mac-device = &wmac;
};
keys {
compatible = "gpio-keys";
reset {
label = "reset";
linux,code = <KEY_RESTART>;
gpios = <&gpio 12 GPIO_ACTIVE_LOW>;
debounce-interval = <60>;
};
};
leds {
compatible = "gpio-leds";
led_system: system {
label = "ts-d084:blue:system";
gpios = <&gpio 0 GPIO_ACTIVE_LOW>;
};
};
};
&spi {
status = "okay";
flash@0 {
compatible = "jedec,spi-nor";
reg = <0>;
spi-max-frequency = <25000000>;
partitions {
compatible = "fixed-partitions";
#address-cells = <1>;
#size-cells = <1>;
uboot: partition@0 {
reg = <0x0 0x20000>;
label = "u-boot";
read-only;
};
firmware: partition@20000 {
compatible = "tplink,firmware";
reg = <0x20000 0x7d0000>;
label = "firmware";
};
art: partition@7f0000 {
reg = <0x7f0000 0x10000>;
label = "art";
read-only;
};
};
};
};
ð0 {
status = "okay";
mtd-mac-address = <&uboot 0x1fc00>;
gmac-config {
device = <&gmac>;
switch-phy-addr-swap = <0>;
switch-phy-swap = <0>;
};
};
ð1 {
status = "okay";
compatible = "syscon", "simple-mfd";
};
&uart {
status = "okay";
};
&usb {
dr_mode = "host";
status = "okay";
};
&usb_phy {
status = "okay";
};
&wmac {
status = "okay";
mtd-cal-data = <&art 0x1000>;
mtd-mac-address = <&uboot 0x1fc00>;
};
|
{
"pile_set_name": "Github"
}
|
United States v. Pohlot
United States v. Pohlot, 827 F.2d 889 (3d Cir. 1987), is a criminal case that summarized diverse uses of the expression "diminished capacity".
References
External links
Category:1987 in United States case law
Category:United States Court of Appeals for the Third Circuit cases
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Modena Weston
Modena Weston (also referred to by her maiden name, Modena McColl; August 3, 1816 in Albany, New York – November 25, 1898 in Bay St. Louis, Mississippi) was a nurse for the Union during the American Civil War.
Nursing career
She enlisted in military service as a nurse in Iowa on September 1, 1861. Her service began with the 3rd Iowa Infantry Hospital. She quickly became a mother to the regiment, working to quell the numerous outbreaks of measles and smallpox. Weston moved to the Benton Barracks, where she was the only woman serving but the station had an adequate supply of sanitary goods. Conditions would not be as good as she moved with the regiment to Mexico, Missouri. Here she established a field hospital where the staff would remain as the regiment moved about. At this location, the water supply was poor and often contributed to further illness.
In a letter to Mary Holland, the most noteworthy moment of Weston's service was during the Battle of Shiloh. Here, in addition to the regiment, Weston cared for the wounded crew of a steamship after the battle. From this location, Weston once again moved to River Landing. Here she cared for hundreds of wounded individuals. Once again, at this location, Weston was the only woman serving. Much like her prior experience, supplies were short and Weston had to request food, and only received supplies after a few days of duty. Weston describes in her letter four women joining her in service, but she was soon once again the only woman as the others were wounded and left. Weston remained at this location until September when the hospital was broken up; after, she served at Jackson until March 1863 only to travel again to Memphis then to Washington, D.C. On April 20, 1863, Weston received a commission from Dorothea Dix, and in January 1864 reported to J.D. Erwin to be sent to serve as a matron at a small pox hospital. Weston finished her military service at this location in October 1866. She was never paid for her service.
References
Category:1816 births
Category:1898 deaths
Category:American Civil War nurses
Category:Female nurses in the American Civil War
Category:People from Albany, New York
Category:Women in the American Civil War
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Hi Adrian, I am a long time home user of Wingate (since v3) and prior to the increase of free 10 user license had a license for the 6 user std version. I am very happy that the increase for free 10 users happened and really appreciate you listening and adapting to the times when *everything* now seems to be wifi enabled, but it seems that the other licensing options may have not kept "up with the times" where they seem quite redundant, for example:WinGate Standard 6 concurrent users $109.95 WinGate Standard 12 concurrent users $199.95
Is there anyone who has a "Standard 6 concurrent user" license (or even 12) when the free version allows 10 users already?
I was looking at the licensing again since I am regularly hitting the 10 "user" limit even with rules in place for things like IP cameras only having LAN access, I think it would be great if you could offer other options for the "Standard" license more in line with the fact that there is already the 10 user free license, i.e. standard 15 user @ $109.95 and standard 25 user @ $199.95And as a side note, it would ease my guilt of using such a great product by being able to upgrade to a useful licensed version rather than using it for free (I am a software developer and know the time and effort put into products like this, and really appreciate your seemingly endless passion to help people!)Just my $0.02
There are still some people who for whatever reason need to purchase the 6 user and even 3 user licenses, for example if they are running WinGate 6, or if they need to extend another license by 3 or 6 users, since the free 10 user license can't increase the license count of any other license.
We did use to have the full size-range of license for Standard, but they really didn't move, once people got to a bigger site that needed a bigger license they tended to want the Pro or Enterprise features, so we removed those SKUs to simplify / reduce the number of licenses.
|
{
"pile_set_name": "Pile-CC"
}
|
Effect of atomic disorder on the magnetic phase separation.
The effect of disorder on the magnetic phase separation between the antiferromagnetic and incommensurate helical [Formula: see text] and [Formula: see text] phases is investigated. The study is based on the quasi-two-dimensional single-band Hubbard model in the presence of atomic disorder (the [Formula: see text] Anderson-Hubbard model). A model of binary alloy disorder is considered, in which the disorder is determined by the difference in energy between the host and impurity atomic levels at a fixed impurity concentration. The problem is solved within the theory of functional integration in static approximation. Magnetic phase diagrams are obtained as functions of the temperature, the number of electrons and impurity concentration with allowance for phase separation. It is shown that for the model parameters chosen, the disorder caused by impurities whose atomic-level energy is greater than that of the host atomic levels, leads to qualitative changes in the phase diagram of the impurity-free system. In the opposite case, only quantitative changes occur. The peculiarities of the effect of disorder on the phase separation regions of the quasi-two-dimensional Hubbard model are discussed.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Spottobrotula
Spottobrotula is a genus of cusk-eels.
Species
There are currently 3 recognized species in this genus:
Spottobrotula mahodadi Cohen & J. G. Nielsen, 1978
Spottobrotula mossambica J. G. Nielsen, Schwarzhans & Uiblein, 2014
Spottobrotula persica J. G. Nielsen, Schwarzhans & Uiblein, 2014
References
Category:Ophidiidae
|
{
"pile_set_name": "Wikipedia (en)"
}
|
In the Original Investigation titled "Estimates of Overall Survival in Patients With Cancer Receiving Different Treatment Regimens: Emulating Hypothetical Target Trials in the Surveillance, Epidemiology, and End Results (SEER)--Medicare Linked Database,"^[@zcx200015r1]^ published March 5, 2020, the authors were in violation of the SEER-Medicare data use agreement and had to remove data from 1 sentence in the Results and remove or replace data in the Supplement. The second sentence of the fifth paragraph of the Results should read as follows: "Compared with 569 individuals in the existing randomized clinical trial,^15^ 940 individuals included in the present analysis were older (median \[range\] age, 74 \[66-93\] years vs 64 \[36-92\] years), more likely to be male (547 \[58%\] vs 298 \[52%\]), and less likely to have undergone prior chemotherapy (53 \[6%\] vs 45 \[8%\])." This article has been corrected.^[@zcx200015r1]^
|
{
"pile_set_name": "PubMed Central"
}
|
Nishant Shah looks beyond the Spectacle Imperative of the Social Web
September 2015. The digital technologies are technologies of watching and witnessing. As the quasi jocular adaptation of an old philosophical conundrum says, "If a tree falls in a lonely forest, and there is nobody to tweet about it, did it really fall?". Facetious as it might sound, there is a way by which we are living in a world of looking. From the ubiquitous cameras on our cell phones to the drones flying in the sky; from surveillance systems monitoring our information flow to the satellites spying on our bedroom windows, there is a culture of watching and being watched that surrounds us. And this surplus of data which needs to be watched and the proliferation of watchers who watch us, have transformed our lives into spectacles – tiny snatches of decontextualized performances which can be stitched together to larger narratives but are largely just held together through algorithms of distribution, storage, and curation.
So persuasive and seductive are these cultures of visuality and practices of testimony, that the digital turn produces a paradox that I label the spectacle imperative. On the one hand, the digital becomes a way by which we see ourselves, see each other, create selfies to ensure that we are seen, and commit to our social media networks, an ever increasing quantified self, where everything from our deepest desire to the internal mechanisms of our body is visualized. Visualization becomes the default metaphor to explain reality, to describe relationships, to tell stories and retell histories, and the quest is to develop penetrative technologies that intersect, interact and investigate the human condition in visual forms. If the Cartesian enlightenment was about producing the human as a thinking being, the digital enlightenment, enabled by fibre optics, is the production of the human as that which can be seen, not by other human beings but by the digital itself.
ITE Algorithm Source: The University of Bremen
We live in unprecedented times where the data that we produce about ourselves is mediated by digital technologies, and at the same time the only way by which this data can be sorted, stored, sifted and made sense of is through the logics and logistics that the digital offers. We might think, for instance, that social media is about connecting ourselves across geographies and lifestyles, with distributed friends and followers, who engage with things we post and share. However, it is the truth that on a social media network, the only thing that sees us, that listens to us – with an intent that in a human being would have been alarming – is a set of algorithms that listen, watch, and remember more patiently and more persistently than our parents, partners or therapists. The spectacle imperative resides in the fact that we live, for the first time in human history, in a condition where the writer of data and the reader of data are both not human. The human is the object that is seen and the site that is always in sight of these digital devices, platforms, and gadgets which become ubiquitous and invisible at the same time.
The paradox of the spectacle imperative is that these digital devices are so omnipresent and naturally a part of our everyday interactions, that they are transparent devices. We see through them. We encounter them through interfaces of intimacy. They vibrate in our pockets and slide sinuously between our fingers. They touch us, feel us, embrace us, but continue to make themselves transparent. Ironically, even as these devices make themselves transparent, they make themselves opaque. The more our devices become smaller, personalized, portable, and non-intrusive, the more they hide from us the mechanisms of their working and the control that they exercise on our everyday life. The spectacle imperative houses and plays this paradox repeatedly. Onn the one hand, we expose ourselves more and more to the digital, creating spectacles of ourselves to algorithms and code that watches and remembers us; on the other hand, it makes the digital itself invisible, hiding it from sight, and thus also making it impossible to understand what is excluded or discounted as noise in the digital networks of life, labour and language.
The spectacle imperative also produces for us a focus on questions of infrastructure, of historicisation, of production, and of distribution. These questions concentrate on how the Internet can serve as either a catalyst for new theatrical practices or a tool to build infrastructure and reach out to new audiences. Simultaneously, only bodies that can become visible, that can bear the scrutiny of the gaze, can make themselves legible and intelligible to the logics of the digital, are considered to be legitimate actors. Thus, we are stitched into narratives of the DIY hacker, of the self-empowered actor, of the change maker who connects with the digital inter webs and starts producing social and political transformation in their immediate environment. As they get mapped on the information terrain, new geographies and contexts produce more stories of actors who take matters in their own hands and produce new structures of governance, participation and collaboration. While there is no denying this wave of individual actors who are leveraging the power of the Internet in order to experiment and innovate with cultures and modes of being, I still want to argue that these actors, who are many in numbers, are at present very similar in their form.
Or in other words, I am suggesting that in the world of big data and the quantified self, where the dominant aesthetic of the Internet is visual, with network snapshots, infographics, and personalized dashboards becoming the prominent interfaces, what we have is a huge scale but very little diversity in the kind of people who get counted and can be presented or visualized as the change actors of contemporary times. Looking at intentions of post-colonial theories, interventions of feminism in science and technology studies, and the impulses of cultural theories of infrastructure and logistics might help us understand that this is not a new phenomenon. And, in fact, one of the problems that the digital turn posits is that despite the newness that it embodies and promises, it seems to continue on tropes of discrimination, making invisible bodies that cannot be presented as spectacular bodies, as bodies worthy of the spectators as well as eligible for becoming spectacles.
The Heathen Without History
In 1835, Thomas Babington Macaulay, whose name can only be explained by the speculation that his parents did not love him enough, in a document titled Minutes on Education argued that "that a single shelf of a good European library was worth the whole native literature of India and Arabia". For Macaulay the native heathen was marked as a subject who has a past, but no history, and hence needs to be educated in English to become the middleware for smooth administration. Macaulay's children, as they are derisively called in India, became the first instance of a modern spectator that was produced by the bastardised technocracy of colonised governance uniting with the codes and codexes of knowledge making. It draws to our attention that the spectator has a much closer relationship with the politics of governance and exclusion than with the processes of literacy and interpretation.
Codes of Domination: Thomas Babington Macaulay um 1855Because instead of importing the books of the great literates from England to India, and teaching them to read and write and appreciate literature, Macaulay introduced two forms in the formal education structures of India – the ledger and the Indian penal code. The surveyors, the accountants, who were keeping count of the wealth and the resources of the British East India Company. The accounting system was not only meant to codify the domination of the British Empire but also to get the heathen counted. The ledger made us intelligible as those who could count and be counted.
What is also present in the construction of the actor of digital technologies, is Macaulay’s commitment to train the Indian native to become a pucca sahib, a true honourable British man. And in the interest of this transformation, Macaulay produced the Indian Penal Code that obsessively sought to regulate the orgiastic practices of the native. He formulated the Unnatural Sexual Acts Law, which criminalised all sexual acts except for those penal-vaginal intercourses that were intended for procreation. It also introduced a series of guidelines on how to curb the homoerotic excesses of the Indian native. One of whichwas to create boundaries in closed spaces, not allowing for the unclean male bodies to touch each other, thus succumbing to the temptation of the flesh. These boundaries became the blue-print of how much space must be maintained between two men in a close working environment, so that they can concentrate on their work, restrain their libidinal desires, and resist the urge to break into song, dance, and sodomy.
Because it was only once the native was taught to count and be counted, to compute and be computed, and once the native was trained to understand the penal implications of his penile desires, that the native could understand the value of literature and the power of poetrythat invited him to wander lonely as a cloud and chance upon a host of daffodils. In order for the native to become an actor and a part of an audience, capable of the acts of reading and interpretation, a massive infrastructure of accounting and policing had to come into being.
This is what I call the dark side of infrastructure. We delude ourselves if we insist that the infrastructure, either for the cultural space of theatre or the diffused space of the Internet, is not only subject to policing but polices the boundaries of the normative and the desired. Anybody who wants to think of either of these two spaces as transgressive, as provocative, as pushing for change, is going to face massive resistance from different powers that wield control and regulate who gets to be active and who gets to be activated in the networked societies that we live in.
Stalker Dolphins
In our imagination of information actors, we presume that the actor is not only human but a particular kind of human – a human who is bothable to receive our information and decode that information. The forms, formats, and tropes of our information production and distribution are severely limited and formulaic because they do not pay attention to the fact that the recipient of information is constructed. The change actor in information societies is taken for granted, and instead of thinking of it as a privilege which is afforded to a few, all the efforts are towards upscaling, mainstreaming, and sustaining the current information packages instead of thinking about how to include new actors into the fold of the information network. For example, the focus of our digital information expansion is on diversity, but it often gets replaced by scale. We often begin with the idea of getting different kinds of users, but end up celebrating having a large number of the same kinds of users. The minute we imagine an actor that is not just a variation on the imagined actor type but a distinctively different actor, we realise the limitation and the possibilities of our current informational forms.
And this trope of a diverse user is perhaps best illustrated through the history of cybernetics and the story of a stalker dolphin. In the 1950s, John C. Lilly was already spending precious research budgets with NASA, trying to imagine what it would mean for human beings to talk to extra-terrestrial life forms. Lilly, for all his fantasies, was postulating a real problem – while we have developed sophisticated information tools to talk between human beings, and modes of translation to talk between human beings who do not share the same language, we haven't really done much work in talking to things that are not human. In his attempt to break the inter-species barrier, he built a 'Dolphinarium' on the US Virgin Islands. It was a house that was flooded with water. Lilly invited Margaret Howe, an animal behavior scientist, to live over land and waterwith a Bottle-nose dolphin called Peter. The hypothesis of the experiment was that dolphins could be taught to identify, mimic and learn human language, and the methodology was for Howe to develop an affective, personal and loving relationship with Peter.
As Howe and Peter lived together, in isolation, spending all their time together in that lab, Peter slowly started getting obsessed with Howe. He started following her, demanding physical contact that can only be described as cuddling, and even sleeping under her bed that was suspended over water. Howe became such an object of affection for Peter that he also refused the company of other female dolphins that were periodically made available to him, and started displaying great physical affection toward Howe. In an attempt to put him at ease, and to reach that threshold where Peter would want to learn the new language, Howe was asked to not only encourage Peter’s stalker moves but to please him – physically and emotionally.
However, by week 5 of this cohabitation, Howe's diary which had begun as a daily account of how she cooked, cleaned and was getting to know Peter, started documenting a new concern. "Peter begins having erections and has them frequently when I play with him" one of her entries reads. Peter had started aggressively stalking and pushing his advances on Howe, so that her feet and legs were covered in minor injuries from his jamming and nibbling. Describing Peter's attempts to 'woo' her, Howe writes, "I stand very still, legs slightly apart, and Peter slides his mouth gently over my shin… Peter is courting me… he has been most persistent and patient… Obviously a sexy business… the mood is very gentle, still and hushed… all movements are slow." She was also scared of where this might lead: "Peter could bite me in two", she wrote. However, unwilling to stop the experiment midway, and also curious about the effects of this intense emotional and physical connection, Howe took matters into her own hands and felt "that the best way of focusing his mind back on his lessons was to relieve his desires herself manually."
The experiment with Howe and Peter has been much critiqued and Lilly's claims that Peter learnt to pronounce words such as 'ball' and 'diamond' because of this new and innovative learning environment are often not taken at face value. While Lilly's claims do not stand up to the measure of scientific and objective truth, and it is dubious whether Howe’s intense inter-species involvement with Peter actually led to the dolphin learning that was hypothesised, there is, in this model, a re-imagination of the learner that is missing from our maps of the actors of information. It is a story of imagining all our current cultural and theatrical practices as flawed, and futile when faced with the idea of a person who does not bear the privileges, affordances, and positions of the imagined actor. The idea of the dolphin as the intended recipient of our communication might offer a completely different form to our visions of change and understanding of what constitutes an intervention. The dolphin, for Lilly, was just a figurehead standing in for an extra-terrestrial encounter in outer space. If we understand the dolphin as that unimaginable change agent who will not be able to make sense of things as they are right now, then we will have to rethink our idea of what constitutes activism. Maybe, like Margaret, we will have to go and pleasure our actors in order to communicate. And our interventions will have to be informed by and travel through these circuits of perversion and paranoia.
Through these two stories from histories of colonization and cybernetics, I have tried to show that we need to shift the focus of our conversations when talking about the actors in our organisations from instrumental relationships about distribution, access, amplification, circulation, etc. I am arguing that there is a more sinister interplay of power and protocol, of algorithms and actors, of networks and performance, that needs to be unpacked in order to think about the political futures and social structures of the entwining of theatre and the Internet. I am proposing that when we concentrate only on the spectacle imperative of change, we forget to look at the larger nexus of technological design, interface intentions, and digital impulses that exclude and discriminate several groups of people who remain invisible in the circuits of change making. I want to conclude by suggesting that when we talk of infrastructure, actions, agents and intentions, we need to move beyond the spectacle imperative that the digital configures us in, and look at bodies, practices and processes that are excluded and kept outside the fold of the change-making ecosystem in order to produce new imaginations that are not arrested by scale and reach, but informed by a commitment to equality, equity and participation.
|
{
"pile_set_name": "Pile-CC"
}
|
Vertical ZnO Nanotube Transistor on a Graphene Film for Flexible Inorganic Electronics.
The bottom-up integration of a 1D-2D hybrid semiconductor nanostructure into a vertical field-effect transistor (VFET) for use in flexible inorganic electronics is reported. Zinc oxide (ZnO) nanotubes on graphene film is used as an example. The VFET is fabricated by growing position- and dimension-controlled single crystal ZnO nanotubes vertically on a large graphene film. The graphene film, which acts as the substrate, provides a bottom electrical contact to the nanotubes. Due to the high quality of the single crystal ZnO nanotubes and the unique 1D device structure, the fabricated VFET exhibits excellent electrical characteristics. For example, it has a small subthreshold swing of 110 mV dec-1 , a high Imax /Imin ratio of 106 , and a transconductance of 170 nS µm-1 . The electrical characteristics of the nanotube VFETs are validated using 3D transport simulations. Furthermore, the nanotube VFETs fabricated on graphene films can be easily transferred onto flexible plastic substrates. The resulting components are reliable, exhibit high performance, and do not degrade significantly during testing.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
This article is more than 4 years old
This article is more than 4 years old
Albania’s opposition Democratic party made a criminal complaint against the prime minister, Edi Rama, on Monday, accusing him of illegally funneling $80,000 to Barack Obama’s re-election campaign in 2012, in return for a photograph of the two men together.
Trump campaign may have broken law by seeking foreign political donations Read more
The Democrats brought the case to the prosecutor general’s office, saying Rama – who was then the opposition leader – paid the money in violation of US and Albanian laws to gain access to a campaign event in San Francisco in October 2012.
Last week, Bilal Shehu, 48, a US citizen of Albanian origin, pleaded guilty in federal court in New Jersey to willfully making foreign contributions and donations to the Obama campaign.
Rama has denied taking part in “any of President Obama’s electoral meetings” or that he paid for tickets to the San Francisco event, although he acknowledged meeting Shehu as “a member of the Albanian community in New York”.
“Edi Rama has committed at least three criminal acts … and personally profited from that,” said the Democratic lawmaker Eduard Halimi after handing over the request.
At the time, Rama proudly used the picture with Obama, months before parliamentary elections brought his Socialist party to power.
“Premier Rama has no connection with any kind of illegal funding,” a statement from Rama’s office said, “and the only ‘crime’ he committed in 2012 is that he had a picture with the US president at a time when the Albanian government and its chief had the doors closed and Albania’s international relations suffered.”
It was not clear if prosecutors would decide to investigate.
Donald Trump blames 'dishonest media' in furor over 'antisemitic' Clinton tweet Read more
According to documents filed in Shehu’s case in the US, investigated by the FBI, he received the money from a foreign source and provided it to a joint fundraising committee to disguise its true origin, so a foreign national could attend a campaign event.
Rama’s name was not mentioned in the case. The documents said the “foreign national” was denied access to the event in San Francisco but was allowed to be photographed with the president.
The New Jersey US attorney’s office said in a statement that no one from the fundraising committee is accused of wrongdoing, and the committee has fully cooperated with the investigation.
Shehu’s sentencing is scheduled for 5 October. The offense carries a maximum sentence of five years and a $250,000 fine.
|
{
"pile_set_name": "OpenWebText2"
}
|
theroguefembot:
As allies, your job is to listen.
As allies, your job is to shut up and listen.
As allies, your job (after the first two things) is speak when spoken to, or to call out other privileged/non-oppressed people to listen to the group you are allied to.
Your job is not to talk over oppressed people.
Your job is not to discuss every single thought about oppressed groups as though you are an expert (cause you watched a single @chescaleigh or @katblaque video.)
Your job is to not make an oppressed groups issues about you. They are not and never will be about you. Thats kind of the point.
Your job is not to make oppressed groups “sound better” or to instruct them to be “less mean.” You get no say in how anger of oppression is dealt with.
In closing, it’s best to think of your allyship as being a house guest. You do not get invited to someone’s house and then rudely demand that they change the decor around that house to suit your tastes. (Nor do you speak over the family who invited you there and demand that they honor your opinion of them over their own opinion of themselves and their house.)
The most you should do, is kindly remind other guests to not be shitty guests and to listen to the host. (Unless asked directly otherwise.)
|
{
"pile_set_name": "OpenWebText2"
}
|
// Copyright 2016 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.
// Bluetooth sockets and messages
package unix
// Bluetooth Protocols
const (
BTPROTO_L2CAP = 0
BTPROTO_HCI = 1
BTPROTO_SCO = 2
BTPROTO_RFCOMM = 3
BTPROTO_BNEP = 4
BTPROTO_CMTP = 5
BTPROTO_HIDP = 6
BTPROTO_AVDTP = 7
)
const (
HCI_CHANNEL_RAW = 0
HCI_CHANNEL_USER = 1
HCI_CHANNEL_MONITOR = 2
HCI_CHANNEL_CONTROL = 3
HCI_CHANNEL_LOGGING = 4
)
// Socketoption Level
const (
SOL_BLUETOOTH = 0x112
SOL_HCI = 0x0
SOL_L2CAP = 0x6
SOL_RFCOMM = 0x12
SOL_SCO = 0x11
)
|
{
"pile_set_name": "Github"
}
|
Q:
Form element name - array type is not working
I have a form with multi-select field:
<select <?php echo esc_attr($item['pgggo_grid_sort_and_filter_multiselect']); ?> name="pgggo-taxon-select[]" class="ui fluid dropdown">
a foreach loop generates this select field but once the form is submitted,
the query is a normal variable instead of generating array variables. It generates the query this way:
?pgggo-taxon-select%5B%5D=14&pgggo-taxon-select%5B%5D=1
Please suggest a solution to have it in array format.
A:
Why the ?pgggo-taxon-select%5B%5D=14&pgggo-taxon-select%5B%5D=1 in the URL
There are three common reasons why would one see that upon submitting the form:
The form's method is post (which corresponds to the HTTP POST method) and either:
a) The current page indeed has that string in the URL. E.g. You're on example.com/my-page/?pgggo-taxon-select%5B%5D=14&pgggo-taxon-select%5B%5D=1
b) The form's action has that string. E.g. <form action="/my-app?pgggo-taxon-select%5B%5D=14&pgggo-taxon-select%5B%5D=1">
Or (most likely in your case), the form's method is get (which corresponds to the HTTP GET method, and get is the default form method).
Okay, but why not ?pgggo-taxon-select[]=14&pgggo-taxon-select[]=1 ?
Because special characters like square brackets ([ and ]) (that have specific meaning) in URLs need to be percent-encoded/URL-encoded (more details on MDN or Wikipedia); hence, those brackets become %5B and %5D respectively in the URL — try urlencode( '[' ) in PHP.
So don't worry.. all is good.
PHP would still treat the input ($_GET['pgggo-taxon-select']) — when sent as pgggo-taxon-select%5B%5D via the GET method — as an array. In fact, PHP (or the server) may not be able to properly parse the input if it was using the unescaped [ and ].
From your comment (formatted for brevity), "but post method does not add any query string but I still get the required array. Can I use a post ? Is it recommended?":
Just use the method that better fits the form's context/purposes — if it's for uploading a file, you should use post; if it's just to read/display posts filtered by certain criteria, you can just use get.
And these resources would help you understand when to use what HTTP method in submitting a form:
HTTP Methods GET vs POST
When should I use GET or POST method? What's the difference between them?
|
{
"pile_set_name": "StackExchange"
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.