text
stringlengths
0
1.17k
[229.38 --> 235.84] understand that not everyone is going to be able to get into every language and pearl during the late
[235.84 --> 243.38] 1990s early 2000s was you know rightfully called the duct tape of the internet and as the market has
[243.38 --> 250.02] grown and there was such a low barrier to entry we have a lot of competitors come in and for any
[250.02 --> 255.68] healthy market of course it's going to shrink our market share and so for a lot of people who didn't
[255.68 --> 261.04] care for the pearl syntax um they were happy to turn away from it so it's kind of sad because it's
[261.04 --> 265.60] a fabulous language and we shouldn't let the punctuation characters in there turn people away
[265.60 --> 271.96] from it but nonetheless you know people have their opinions and just as i understand that i also like
[271.96 --> 277.36] other diverse languages i i enjoy python i enjoy ruby i i like prologue i think it's fascinating
[277.36 --> 282.76] uh but i understand not everyone likes everything uh lisp i think is phenomenal language but
[282.76 --> 287.96] i don't care to program it just because i just find it so frustrating and ugly when i play around
[287.96 --> 293.08] with it no no offense to lisp programmers out there so i i don't really stress about it much i understand
[293.08 --> 299.26] that it's not everyone's cup of tea so you've been doing pearl a long time you've written a book on
[299.26 --> 305.32] pearl you have a consulting company that consults on pearl uh what is it about the language that you fell
[305.32 --> 314.08] in love with it was an accident um in i think it was 1999 i was a mainframe programmer and i was
[314.08 --> 321.72] working i was mostly doing cobalt development and one of cobalt's worst strengths is worst abilities is
[321.72 --> 327.00] dealing with freeform text and that's pretty much all the web is incidentally which is why cobalt even
[327.00 --> 331.00] though it's tried to it's never broken up the web and there was i was working on a program to
[331.00 --> 338.60] convert nt csv files into the mainframe fixed width format that cobalt is really comfortable with
[338.60 --> 343.96] and it was about 150 lines of code and there was a bug and someone didn't understand something called
[343.96 --> 349.66] the unstring function something we would call split in many modern languages which splits a string on a
[349.66 --> 355.34] character and i fixed it i got it down to i think like 80 lines of code and this unix sysadmin kept
[355.34 --> 361.46] telling me you got to check out pearl so i checked it out and i got this 80 line cobalt function down
[361.46 --> 367.12] to about 10 lines of pearl and that was with error checking and it was actually fairly readable and i
[367.12 --> 371.96] was thinking my goodness what the heck am i doing and when he eventually left to form his own company
[371.96 --> 378.32] he said come along i know you can do this and i haven't looked back though the ironic thing is i enjoy
[378.32 --> 384.96] a lot of other programming languages i program in c assembler variants of basic java ruby python
[384.96 --> 392.18] but i joined pearl right at the time of the dot-com collapse so i stuck around with pearl for so long
[392.18 --> 397.98] that after the economy rebounded i found myself in a situation where people would see so much pearl
[397.98 --> 402.80] on my cv either they didn't want to offer me a position or they would offer me a junior programmer
[402.80 --> 408.54] salary so i wound up sticking with pearl and i've been specializing in it for about 15 years now
[408.54 --> 414.90] it seems fitting that the the thing that brought you to pearl it seems to be its its best trait which
[414.90 --> 420.56] is and what it was designed for right isn't it all about text extraction and manipulation that's
[420.56 --> 427.12] initially what was going on larry wall uh the creator of pearl uh he originally released it in 1987
[427.12 --> 434.70] and he was trying to handle a lot of problems that said and awk and other tools were supposed to be
[434.70 --> 440.28] doing but he wanted to do it in one tool to make it very easy i believe for reporting for nasa as i
[440.28 --> 447.02] recall and then he eventually released pearl uh open source to the community pearl 1 in 87 and it just
[447.02 --> 454.30] took off from there it was just so phenomenally easy to hack and pearl i mean today i often find myself
[454.30 --> 458.14] writing quick bash scripts and as soon as they start to get complicated i say okay forget about
[458.14 --> 463.86] this and i switch over to pearl because it makes things so easy but at the same time i also specialize
[463.86 --> 470.72] in extremely large scale uh websites you know database driven uh that use pearl almost exclusively
[470.72 --> 476.26] as the back end so it's everything from the really tall blue really small blue things that we have
[476.26 --> 480.96] to the very large scale websites some of the largest e-commerce platforms in the world are driven
[480.96 --> 487.76] with pearl and it's just amazing how easy it is to shift back and forth like in java i'm not going
[487.76 --> 493.66] to use java to hack out a small uh small utility for gluing things together it just wouldn't make any
[493.66 --> 499.30] sense at the same time uh tickle might be great for you know a lot of smaller tools but a lot of people
[499.30 --> 505.22] complain it doesn't scale as well though again no offense to the tickle community um so it's it just
[505.22 --> 510.92] really feels fills a sweet spot for me of being able to solve virtually all of what i tend to do
[510.92 --> 518.58] on a daily basis one of the uh the biggest pearl advocates and fans that i know on the podcast scene
[518.58 --> 523.94] is john saracusa who um writes pearl you know professionally to this day and loves the language
[523.94 --> 528.74] and one of the things that he says that's interesting and maybe you can tell me if this uh resonates with
[528.74 --> 535.82] you is that pearl is really kind of a formalization of the unix way and kind of uh taking those ideas
[535.82 --> 541.00] of those small tools and those command line tools and wrapping them in kind of a nicer uh language
[541.00 --> 546.62] is that does that resonate with you or yes it does and i'm actually going to go back to cobalt for just
[546.62 --> 550.84] a moment if you don't mind there's a reason for that so one of the things fascinating about cobalt
[550.84 --> 557.22] what made cobalt so powerful and why it stuck along stuck around for so long is because cobalt's not very
[557.22 --> 561.80] good it's not very powerful it's hard to write big systems in cobalt so what you do is you write
[561.80 --> 566.26] a small cobalt utility which maybe reads some records from an isam database and stores them
[566.26 --> 572.40] in the file but you have jcl job controlling which kind of it's tough to describe it doesn't really
[572.40 --> 576.72] have a good analogy today but jcl would have different steps so you call a step which would
[576.72 --> 581.62] read that a cobalt program which read the data saved to a file you call another step which would sort
[581.62 --> 586.76] that file and then the next step might load another program which would read that file
[586.76 --> 592.96] add some more data in save it and then you call another one which would take that saved file
[592.96 --> 598.64] pass it on to another system basically it was a unix pipeline and that's part of what made cobalt
[598.64 --> 603.88] so incredibly powerful because it wasn't powerful so people built a lot of small decoupled tools
[603.88 --> 610.66] and kind of piped them together with jcl so that worked out very well for me when i was transitioning
[610.66 --> 616.24] into pearl initially and getting used to the unix model because it was used to the way my mind
[616.24 --> 623.36] already worked build small tools pipe them together so that's part of the reason why yes i write a lot
[623.36 --> 628.72] of bash i write a lot of small bash utilities to get stuff done but anytime it starts to become
[628.72 --> 634.10] painful and bash and anyone who's done enough bash scripting knows what i mean i just switch to pearl
[634.10 --> 639.80] and i can do the same thing and it winds up being it's not quite as simple as bash but once you get
[639.80 --> 644.06] to the stuff that bash is you know a little bit weaker on or maybe my bash knowledge isn't as good
[644.06 --> 649.88] in pearl just makes it so easy to glue all these different tools together to shell out to some of
[649.88 --> 655.60] their program fetch its results uh you know fork off multiple processes run a whole bunch of stuff
[655.60 --> 661.80] aggregate them together and push it out there it's just it's lovely it's simple and you know from
[661.80 --> 667.12] scaling down to that small scale the really tiny things you do up to the big large scale systems
[667.12 --> 673.66] it's just it's always amazed me how seamlessly it tends to do that let's talk about some of those
[673.66 --> 678.50] the large scale systems you speak of um do you know any off the top of your head that are like
[678.50 --> 682.24] you know well-known sites that people may not realize are actually powered by pearl in the back
[682.24 --> 686.08] end uh depends upon what other people would think of as a well-known site so i live in europe
[686.08 --> 694.18] and one of the well-known sites over here is booking.com uh until the ipo of alibaba they were the third
[694.18 --> 699.98] largest e-commerce site in the world after amazon and ebay i mean they're huge they're not as well
[699.98 --> 704.78] known in the united states but basically they're an online hotel reservation system and they're massive
[704.78 --> 710.44] and yet almost the entire back end is written in pearl and i remember when i was working for them
[710.44 --> 716.56] one of my first days there i was walking by this guy and he was hacking on some java and i was surprised
[716.56 --> 721.24] and i said what are you doing java programming what do we do with java here and he said well we don't
[721.24 --> 725.44] we're taking all of our java programs and we're converting them to pearl just because it's easier
[725.44 --> 731.58] to work with which i found rather ironic because sometimes you hear about it going the other way
[731.58 --> 735.24] around people are converting pearl some of the language and here they're converting from some of
[735.24 --> 739.82] their language into pearl and it's something very common for them but they just found pearl so easy to
[739.82 --> 747.18] work with so that's possibly the biggest uh company i know of i work for the bbc also um world's largest
[747.18 --> 753.98] broadcaster they had 26 000 uh people when i was there and i was working on the central metadata
[753.98 --> 759.50] repository which basically that was information about you know what their schedules were what
[759.50 --> 765.34] programs were on telly and i found it rather ironic that me an american who didn't watch tv was telling
[765.34 --> 771.10] the british people what tv they were going to watch and all of that was managed through their pip system
[771.10 --> 779.46] all written entirely in pearl and just many many companies like that crowd tilt now known as tilt.com
[779.46 --> 786.26] which is a popular crowdsourcing system is written entirely in pearl there's actually an mmorpg called
[786.26 --> 792.74] lacuna expanse which has been written in pearl lots of large-scale systems some are well-known some are less
[792.74 --> 798.98] well so there's a lot of it out there um yes man it's been a long around a long time there's a lot of it
[798.98 --> 803.92] out there it has a lot of virtues what is it about it it seems like it's behind it seems like pearl's
[803.92 --> 810.24] behind the scenes is it just bad marketing or um is it just communities that you know don't necessarily