Keynote: Welcome Inside the Head of Larry Wall - Joe Armstrong and Larry Wall

okay well you know parole has always come at it from net the language thing from a natural language point of view because I was trained more as a linguist than a computer scientist and some people would say it shows and but you know with the natural language you learn it as you go you're not expected to know the whole language at once it's okay to have dialects the gold ESL's if you want other things you know languages natural languages evolve over time and they don't have arbitrary limits and they naturally cover multiple paradigms there are external influences on style you write poetry difference and then you swear I hope it has fractional dimensionality to it easy things should be easy hard things to be possible and you know if you get really good at it then you must be Kampai so that's natural language but Perl has always been sort of the same kind of idea makes them easy things easy for most the people but you don't have to be a computer scientist to understand it if it's accessible for mere mortals to begin with but it hopefully doesn't run out of steam later so a little background in the year 2000 we said maybe it's time to break backward compatibility what just once and see maybe we can afford to do that get off of the worse is worse cycle crank the thing once for a worse is better cycle so if you could break one thing what would it be I expected maybe 26 suggestions we got 361 another thing that happened about that time was Paul Graham wrote his essay is a hundred-year language and he speculated on what what the characteristics of a language that might last 100 years are and you know it has to do with whether it can be extensible whether it can evolve over time gracefully and so we put a lot of these ideas together and thought really hard and so we came up with a whole bunch of principles in the last 15 years there's no most important principle trip this one we want to torment the implement on behalf of users if the error message is less than awesome we can set it at a bug every glim comes with corresponding lat put that in the corner can capsule out your cleverness and then we use the heck out of it so all of our pattern matching what we call smart matching is all in one one table and every type knows how to do with smart matching and it's that way consistently throughout the language the press knows tables another place where there's a lot of cleverness encapsulated in one spot but you just learned a is a funny kind of a be but call it a that's a lot and we learned from Perl 5 I won't go into that give the user enough rope to shoot themselves in the foot but hide the rope in the corner ok every feature should have high leverage but not too high pick the right default this time around that pearl size cut sort of refrained from picking defaults for object orientation and that turned out the mistakes so ya better default and everything should be strangely consistent so kind of always things but the thing we really came up with was you know girl as the culture is is postmodern but there really is no one true language not even Perl 6 of it is the one true language because Perl 6 itself is a braid of sub languages planks for short convenient and they interact with each other and you can modify each each part of the braid so here is almost getting long enough to braid now I never grew my hair out before about to try that but a braid of languages let me illustrate ok one of the braids of languages is for each of the different slots that an operator can be so we can redefine or define not redefine a postfix bang operator we don't happen to have it built-in wonder of wonders so if I uh if I defined that municipal should look almost like Erlang you recursive function obviously but you notice that over here we're actually using the new symbol in part of its definition the the way the flying works is as soon as it sees that name of a new function it derives a new parser with with a mixed in rule for parsing that it automatically writes a new lecture for you so you never have to write your own lecture a righteous on its own longest Turkic token matching and it's all already available by then so if I run this here it actually comes up with them yeah yeah it we also have obviously long big integers by default so here we have same thing now a lot of people are allergic to schedules we as in protocol tree culture we like our schedules but you can write without them so here we have the sigil list style here where we just called our parameter n and that without that works also or we can talk directly to the dispatcher we've unified functional programming and object-oriented programming they're just all functions with different dispatchers so here we can actually talk to the dispatcher and say well call the same function but with this argument and that set will also work but that's not how we usually write we don't usually write tail recursive things in perl 6 you'd usually write a reduction operator this just says multiply everything from 1 to n and that works too and there's a little philosophical thing there is that even though we're building a lot of functional programming capabilities into Perl 6 this whole head tail recursive definition of things tends to work against concurrency we're more of the view of guy steel and fortress where we want to be able to subdivide things equally list so our list they're higher abstraction levels than necessarily the head and tail of Lisp so going on 15 years later what have we got well we we got in 2015 we got our first real release of Perl 6 so the version number is six dot C Christmas the next one will be six dot d Diwali but what what do we what do we have we have a fun handy scaleable learning Bowl expressive optimizable maintainable they just keep getting longer and longer don't they multi-paradigm gradually type Unicode friendly concurrency aware test suite specified functionally reactive parametrically goes a Viking safe linguistically malleable multiple virtual machine meta object protocol defined representational e polymorphic that's important and every one of those things was worth the talk but you know pearl sixth we want it to be the last programming language you'd ever want to use no that's not right now we wanted to not run out of scheme when you when you need it not to okay so but we've borrowed a lot of ideas Haskell we we stole I mean borrowed the you know idea of laziness and some other ideas from small talk and and Java Smalltalk we borrowed the idea of trait we call them roles instead but we instead of just doing small talk traits for compile-time mix-ins we also added in ideas of interfaces and mix-ins and and generics so those that one construct serves all of those kind of things as an immutable reusable bits of code that are not classes borrowed some things can go have promises and channels so if you're familiar with those older functional reactive programming from c-sharp and I guess their line we have borrowed some ideas from from APL indirectly not the syntax of course but for instance we have a lot of operators a lot of list operators but we can compose the meta operators and this is kind of an idea from APL here's just a plain old cross operator takes two lists if you run it it produces the Cartesian cross-product of those well but what if you add a concatenation operator there it's crossed with so be like zip with cross list etc so this this will instead do a cross concatenation of everything and that's just one of the meta operators that we have here's something that's a kind of fun functional programming I mean it says I have a sequence and here's 1 2 3 you figure out the rest of it and it does but again Haskell pleasure with so was it that I'm smarter than you I'm not going to try to do that or it doesn't matter if you like it you know that that's by the way is red one two three dot whatever the stars is usually red whatever and what it really needs is delegate to whatever you're calling to figure out what the appropriate thing there's but you can you can have the explicit infinity you can write it in Unicode if you want network to do but what if you what if you actually just want to have part of it well you can you can slice the subscript all of these freaking the same things we use zero up to but not including so often as a range that we have a unary operator that does that so all of these will produce the same sequence sequences also know how to iterate on on strings so you can do that now you guys have smartphones and they're they're always fixing your quotes to curly quotes for you and that's frustrating right well you know what if you actually type that into it doesn't care curly quotes are fine how many you from Europe how many you have different curly quotes yeah well some of some of you had these right at work and some of you have the closing curly quotes the other direction that also works and what if it actually turned to your dot dot dot into a Unicode character well laying out that work too so we're very very up on Unicode here we process everything in terms of graphemes flex lift only better and we call that normalization form graphing but basically the idea is you take any composed character whether or not it has precomposed a variant in in unicode will pretend like they're all free composed so those are graphing it's the native speaker or you hear it physicist he wants vector move and he wants to think think of that as a single character or she well you know it is one character in Perl 6 it's two code points but it's one character it doesn't matter how many markings you have on if you in Perl 5 if you say this it says that's 49 characters oh look I forgot to tell if it was utf-8 so we say utf-8 it's still 27 characters Perl 6 how many is it it's five if you have to watch the right your your numbers in Tibetan I work or kumea form this it all falls out of a Unicode I mean they just apply this so it's trivial to hook it up really vulgar fractions they work you might think well that's a floating-point number no it's not it's a rat has a nice mellow one and what are the rats methods well has lots of rat including attorney to a fat rat to be heated enough okay well I said one of those messes is nude numerator and denominator but you'll never forget this method ninja there's got to be a joke about nude rats in here but I don't know what it is or you can get the pearl representation of it you see it's sort of going back and forth here between functional and you can treat everything as an object much like in Ruby and this is an exact perl representation of self at values to run period okay well what happens if you speaking of wrath what happens we say this in in pearl high if you try this new favorite language it probably is going to say these are not equal but in in Pearl fixed we see the yes it is because those are not floating point numbers in Perl 6 those are rational you actually have to put the e 0 on it if you want to want to do things in floating point okay so since we do things exactly in rationals does this exactly match this end point of course it does try that in your language and it might not match it might just keep going on forever okay but you we can say okay here's one that does go on forever but give me the first forty values of it so you see it into a geometric sequence if and here's another geometric sequence that just happens to have a multiplier of minus one so it can do that one two and yes you can just say two to the 32 if you want alright what we'll do with this looks like it wants to be a Fibonacci sequence doesn't it well we're not that crazy if you really want a Fibonacci sequence then you can either name these the built in in fix plus operator that's its actual name or you can write lambda in any of those three ways and those all produce Fibonacci so you know we had a earlier talk where we had had Fibonacci a 40 in a cook couple couple seconds to do it it takes point zero five seconds in Perl 6 to get the fortieth Fibonacci but of course we're not seeing it certainly more fun function composition okay log of a - number is quartz going to blow up but if we compose it with absolute then we can get a safer version fold operators you can write reductions in the shorthand that we have but there's also the longer form if you need to get fancier so these all add one through four quick what's the sum of all these numbers right it's cheating of course it's optimizing that to a sum method on a range object which it knows how to skip over the intermediate value here's a scan if you in the half terminology so so that gives you the half scale numbers triangular because your numbers or your structures tend to get larger and larger so there's sort of a metaphorical triangle there so the fur looks like typing the wrong thing okay so there's the first 40 triangular numbers there roll 5.5 Yahtzee dice here and print them out unfortunately Unicode doesn't actually give the numeric values to little dice the last and usually original ray but for those those dice okay the sum is 15 mins 1 the range is all the way from 1 to 6 and the least common multiple is 12 because that's an index operator we can use it as any reduction see more things to look at don't have to specify elf over our fold because if you do with that from the operator self so it knows that exponentiation falls from one direction and multiplication from the other automatic identity if you if you like to do a large factorial that's one thing but what if there's only one thing we'll go get one well is nothing in that so still one because it knows that that's the identity out if you reduce on a comparison operators and it I made these all variants of the end key and that was a mistake control and is not the famous okay okay so I guess I get true and false there the identity value for that one is true identity for men and maxes plus and minus infinity of course GCD I don't know one whose there are demeaning 40 cedis okay other meta op we saw a text earlier suppose we define our own our own data operator here which puts a bow tie so that puts puts the arguments in Reverse this does as if so those in parallel zip with basically the X we already saw does the cross you can stack them so the reverse X and yeah you could you could have a operator this you know 500 characters long if you want to do no arbitrary limits there's X reverse which you know it's actually different if you do reverse X reverse it comes back up in the same thing so so just to sum up quickly we do a lot of stuff with concurrency now in perl 6 that will look very familiar to go in Erlang programming that the structure of the scheduler threads versus the witch cult process is quite similar reactive programming is these new things that work where you can write things that look like loops but they do the right thing and compose with reactive town responses the things and this does kind of stop on this one one of the important things that we do now is called representational polymorphism and that means that even even your classes in perl 6 don't actually care what representation is so you write your class the same way your message the same way for accessors and it doesn't care whether it's the built in opaque type or whether it's some DM array of underneath or a fee structure or C++ or or Python objects or Ruby objects those can all look the same so that makes it very easy to interoperate with any any other languages that we can talk to that's probably enough probably too much to 300 years time all these programmers you have to read legacy Erlang and segment early and they go what it's like it'll be like the ancient hieroglyphics a nobody was dared change it because it all works people who do this to us fortunately I'll be dead right yes and also if you design language with you in this course you get no credit for all the programs at work but your program to speed written in your language which failed at your hospital of my course yes one of the things we started out with the window when we decided to break things and come up with Perl 6 was you know we already know what our old mistakes are let's let's let's throw those out and make new mistakes this time but kind of related mistakes in the sense that pearl as a overall philosophy of how you how you think of programming as an activity embedded in a culture and that the culture says a lot to how you actually structure the style of the program and you know externalities will often determine whether you want to be object-oriented or functional or do logic programming or what have you just as in a natural language you know if you take a class in in poetry you're going to produce you know work the literature that hopefully look and work like poems but if you take a math class and write a math paper it is structured more like a math paper and but in each case in an all sort of literary work it's this negotiation that is happening between the writer and the reader whether that reader is a computer or the next programmer you know who's going to read your code and you know as an actor you know there's this negotiation of here is I'm starting off in sort of with a an understanding of my character lessons here but I I switched into more and more specifics about exactly what language I am I am programming it and the big idea in Perl 6 is that unlike Perl 5 which was always really confused about what language it was parsing and so don't ever try reading its parser you'll go mad Perl 6 in contrast that's one of the most things we decided we liver make again Perl 6 always knows exactly what language is parsing and a strict one cast parsing like Pascal and at every point it can drop in when you define a new operator or you pull in a new slang or if you say use COBOL it just drops into that different language and it might or might not be really related to the outer language it drops into the parser for that language as a pure parker and that knows what to do with that language and then at the end of that block or program the sub program it comes back out and it restores back to the previous lexical scope the lexical scopes are very important for knowing exactly what language you are in and we never you know cheat by looking ahead and seeing what classes are declare don't even know how what the class declaration will look like at the end of the program necessarily it might be we might be in a completely different program so we don't cheat anywhere in Perl 6 so it's very clean linguistically and this buys us a lot of power that we think will eventually control that that divergence that happens to all languages you know either your language is too brittle and people just invent different languages let's do the same thing with different syntax or or your languages is too mushy and and it just becomes a mess kind of like some some Perl 5 is unless you're disciplined so to avoid either of those big mistakes is we have this just as as Erlang is very strict about its process model we are very strict about our linguistic model and so we think that that that will help us over the long term to be more like what Paul Graham was talking about when he talks about 100-year language sorry its preparations are telling us conversation together I was looking on the internet it's quote to do or something I think so that oh wow well now I'm say anything no but and then I discovered it that you want you're older than me so so so tell you I'm not sure that I'm a year wife so much more with my container and as wiser than analysts attempt but it meant that we've been through the same shift in generation 15 hardware and what what was always kind of coupling these is where did we go wrong because and I shouldn't either in my mind certain certain dimensions of computer that are fixed in my memories so one limits that took a long time to break with this 1/4 or 1.4 for megabyte limit of a floppy disk yeah and for a long time development increased to the point where you got to one point four four megabytes and then you stopped I just melted limit the man when I was first sending our programs Usenet articles could only be 30 32 K characters right so if you wanted to send out software then you had to have a little program that would split your program up into separate files that could then be me I remember Tom's Tom's recovery disk you know we had a complete UNIX kernel and everything on a 1.4 4 megabytes floppy disk and I remember the first from Airlines we did get essential they fitted into 1.44 megabytes and then things started to change and these went away and suddenly you're not aware of that anymore and you go into the future and you look back and suddenly said what the perception that an operating system is you know a small operating system images that you gigabytes or something like this so where do we go wrong well we're not we're not wrong we're just displaced because you know yeah that our normal machines are getting bigger and we're powerful but you know this whole internet of things we're gonna you know you're not going to put a you know a huge computers I delightful we can put it right now I'm good comics I want into it to run it rather CPU hot enough it's just being consistently good but it's going to be a mix pod more powerful in the clay probably and you know that that's part of why I should not worry too much about making a large language in Perl 6 obviously the larger your language the harder it is to support we do have a portability portability layer of a restricted subset that does make it easier but you know one one worries a little bit about making it too hard to put onto different VMs we currently run on our own VM what's called more VM historically used to run on parrot VM we have been on c-sharp and on on Haskell those are all prototypes that has sort of gone away we also run on the JVM currently so we've got a story about portability but under what but I'm mo um I'm more thinking in when we have megahertz clock yeah you know why why are why isn't an operating system booting them and generally microseconds because yeah because well things thing he espies well then where it's Alan Kay cat we should you have to reinvent Gator you know you know computers are getting more and more like people and so the principle is applied to people are starting to apply to computers and this principle is every computer will be promoted to its levels and compensating okay and you know artificial intelligence is a starting going to start looking like natural stupidity so I'm not sure that you know even if the singularity comes well I think it's probably a plurality not a singularity but you know I think the the the foibles that we have as as humans yeah that the computers will just make the same mistakes faster unethical to others affected if you go back to the mid eighties but programmers could talk to each other because everybody even if they knew different programming everybody knew see everybody knew may kept repeating you bash and the matter they can't talk to each other know what builds they all are using goo BR go beyond grunting you know my grunts my grunt Guffey's will yeah I so it'll be yeah a lot of coming in here I said this OTP still mean what it used to be but yeah you know we are we are in the process of becoming fossils and the new culture will always be built up on the strata of the old so it's a healthy process that that the new generations will will take what they can from us all teasers and you know take what they like and leave behind some of the things that maybe they don't like so much but I mean now I think when we started project we didn't say for you you know you where you're learning programming you had the choice of learning assembler COBOL or Fortran I learned basic oh we had machines with that advanced it's actually my first the computer program was on a programmable calculator but not even I think naysayer or some some off-brand is you could program it 120 steps that were like compresses and I almost got sick Tac Toe in to 126 I'm sure it was a very stupid program and if I could do it now I probably be better but no actually my first computer language was a basic plus on a-deck Christus and she the one lemon table uh my in 1972 it was actually on a PDP 11 20 which was a forerunner to the PDP 11:45 okay have you hadn't done any programming to pull it out it's only on the only in my last quarter of high school that when we got the few calculators and I was I was supposed to be doing independent study in math and most of my inside the calculator so I started with Fortran in 1968 I think and with a three week turnaround times of your programs because the first week you you get it on the coding forms and I said too often and I did have to do COBOL in a bad system we have to take our card decks down and of course I never got to clean the pile in but on the other hand you know I didn't really study all that much instead the night before the final I read the entire IBM COBOL manual and aced the test but yeah there's lots of people people talk about so well Coco has has had a good run and I do have to you know sometimes disabuse people of the notion that you know if a programming language is like a networking like a natural language they really think of COBOL which sort of Cargill cultured phrases in without really doing anything and it was really just a procedural language that happened to have English phrases with key words but it didn't really work like the natural things on a deep level well I I knew I knew COBOL was not for me as soon as I I saw you called you call a subroutine or you thought you say compute you say what it is you don't even call you why don't you do a subroutine that's two letters you have to even both I forgot - yeah there's like a seven-letter word for for calling a subroutine I said why why are they using long words for common stuff since the beginning of this inkling that you know it's going to use something frequently it should get rubbed down and be the short thing the the term we use nowadays in InDesign is Huffman coding now Huffman coding is a bit encoding that we're we're frequent characters get a small number of bits and less frequent characters get many bits but the metaphorical meaning of that is you know make the common cases short and this happens in in natural language too if you look at almost any natural language the most irregular parts of the language will be the most common these verbs amount and that's because that is where you you use the most you need the most irregularity there for information density and clarity and redundancy of communication so that it's not really surprising this is the most irregular parts of natural languages are are the most frequently used end of the most rubdown bits of it so similar things can happen if I'm also under get the at what point in time did you I mean did you at some point in contact I know what I'll do I'll make a new programming language all the other I never did - I did my first real language well I did a language for a compiler class with the Dragon book yeah yeah I'm mostly ignore all dragon back I also ignored the dragon book and then use yes but that's okay I'll a whole and I are friends now yeah he's a great guy and but you know I one toy language event then I had this problem that I was programming for our school our computer center in basic plus doing real work for the college and and I needed to be able to sort of link in various basic routine so I need to kind of an inside-out language nowadays for an inside-out long as you think of something like PHP where you know it's mostly text but then there's sort of code interspersed with the tech this was a different language but you could interspersed lines that that would control how the text were used so it could basically selectively link in different routines and this was called Jam which was the jury Pig jury jury rig all-purpose metal language hi jammers Jose abstract machine yeah 32 yeah but that was that was actually my first compiler that was that was just a recursive different kind of kind of thing and but but about stages could scratch your programming ages oh yeah yeah oh well that people can use yeah and but then I went to work in actual industry after after some grad school and and had to do a compiler in Pascal for for a discrete event simulation and I you know I've worked some macro systems for the RN program that I had written and so you know I sort of had this general feeling that I could do this sort of thing and I I'd published things before the RN program the patch program which does you know certain amount of parsing so I already had this kind of inkling that you know I like writing compilers and what if I like something other people might like it so when when I sort of had this realization that the UNIX had this gaping hole you could do shell programming which was you know very you could whip things up I call it whip up the tube and you had C programming which was very intricate and you could do anything I call that manipulate city if you if you grasp if you think of that as you know either or then you you know programming one of the other but I'd realized there's two different dimensions and it opened up as kind of a two-dimensional graph and I realize it's a big hole out there where we're something that should be you could whip stuff up in quickly like shell but that could do more intricate things than shell programming to do and and work more like a actual programming language even though that kind of thumb was a violation of UNIX philosophy and do one thing and do it well it actually ended up being a toolbox such that people could then make other tool and kind of fit back into this little box so you know I was already on my you know my rebellious slide into changing the world at that point and so you know not not everyone in the UNIX community appreciated that that you know I think Henry Spencer famously called the early parole office skin cancer IIIi think I was ugly you know I got coded transmission sun-god well you know but it's uh it can be written to look like that yes with this upon even even even pearl five pin can be made today I'm and reading your poem book and thinking yeah I like the gate with pearl six yield you'll discovery to that you can actually write some beautiful codes with it but I was kind of interjected by the way you started off with a I think it was a character count or a word counter or something like that you wrote it more or less as I would understand and see or something the while loop and getting tokens and things always we can omit these brackets that we can remove these organisms and the program just got shorter and shorter children up yeah well I I did a couple years when the upper skated C code country I saw that yes what what did that produce do Oh one of them was a Dvorak keyboard emulator so ye you like ran it and and started typing in and translated all the characters the other was a Roman numeral calculator that was all what like was all written in roman numeral right so these techniques you use with you into the source kind of problem when I started working on on Perl 6 and we as a community we were starting to decide how to divvy up all the work the people who were with it said that we will implement Perl 6 you stay out of the implementation you are just the language designers because we see what the insides of Perl 5 look like and we don't want the inside of Perl 6 to look like which is funny because it sort of ended up doing the same thing you know you can't program the sea without writing a bunch of macros did you keep out of it will you put I can actually keep out of it and that's part of the reason that we we had I suppose you know three or four prototypes that didn't really work they say we learned everything even the failures though we learned a great deal from each of them the each person who tried to implement Perl six and had different ideas about it they they taught us you know different ways to think about things the first implementation was actually in Haskell that really tightened up a lot of our thinking about how did you remove a lot of stuff as well as adding stuff I mean from the ideas come from Perl well no I mean from poke successful as a designer you say ah I think we'll do that and we'll do that um and it just becomes so complicated to implement that you remove it or uh it was kind of careful not to put in things in the first place that we didn't think would be quite useful what we did discover along the way was that many things that seemed like separate things on the surface ended up unifying underneath so it turned out to be you know really easy to to get most of functional programming and most of object-oriented you so if i Nick as we have said you know every time you add something to program and you take something out you build it so - yeah that remains for you I freely admit that you are more amenable than I am so I keep wanting to take eggs a solvent or not and I need I need the minimalist running alongside my chariot saying you're mortal and so I have a great appreciation for for the idea that of trying to keep things simple as simple as possible where I uh where I think it sometimes goes wrong is you know the full phrases keep it as simple as possible but no simpler and real life is not simple and the problems that are presented to computer science to solve some of them are you know they're math problems they work great in functional programming some of them are simulation problems and those tend to work better with with object orientation logic whatever whatever your whatever you you slice up your your paradigm you know it's quick and dirty scripting so that there are there are languages that excel in each of these areas and one can achieve a certain kind of purity I suppose even in the PHP where the PHP realm there's some kind of purity as they are achieving but when you start looking at a language that wants to encompass the whole world like like English it doesn't have to be perfect at everything but if you can do 90% of this paradigm and 90% of that paradigm and get those talking to each other underneath in a smooth fashion that's what we think we would achieve with Perl 6 if you uh if you scratch a method it's really just a function just has a different dispatcher that says how about of how to find it and call it yeah I think fine chess guys walk in and I think there are drinks afterwards and ok and we could actually see her talking Dresner we could we could um so we could yeah let me just have one pile in Soviet that's the main the main thing that I think that we've really got in from functional programming is the idea of a coherent type system Perl 5 had an incoherent type system and and so the complete the computer was often just as confused as the programmer is the type of thing with the Perl 6 we don't we don't really have types for the benefits benefit of strong typing it's more like that's that third place second place is we can optimize that the first place is actually if therefore multiple dispatch and pattern matching could use what more you can and this is pervasive pattern matching and multiple dispatch whether it's two methods or functions I mean you saw it there with there's a perfect example is built built in at a fundamental level and that that's why Perl 6 love types is to drive that and it actually once once you've uh once you've done your multiple dispatch and decided which thing to call you don't even have to check your function arguments you've already checked your function arguments to do this you just patch so it in a sense dynamic type checking comes for free as soon as you've done that so that's that is another thing that I want to thank the the functional programming to me for for letting us feel but in some sense you know we have our separate communities but we're all you know one big community in world was trying to get similar things done and you know going back to anthropology there's there's people who live in one community and you know they tell the stories of that community and there's a tribal leaders there but there's also the merchants that go back and forth between different community and healthy cultures need both of those things so it actually pleases me to see you know Erlang and elixir sharing a conference like that pearl has shared with the Ruby community over the years especially in Japan yeah because of coke we can build systems with one language yeah and so you know our slogan there's more than one way to do it that's lies outside of pearl as well as inside so yeah I think it was great and thank you for inviting me and thank you [Applause] [Applause] any kvetching where's wimp appear well of people metaprogramming it means it needs a champion maybe that will tickle one of you guys we definitely have tried to you know make it possible to put put our language on two different vm and it does run on JVM and JavaScript is nearly done if you can put it on JavaScript you call it anyway [Applause]

Channel: Erlang Solutions

Views: 6,672

Rating: 5 out of 5

Keywords:

Id: SpH9BTMZKXc

Length: 51min 41sec (3101 seconds)

Published: Fri Mar 24 2017