Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    S Wade
    @swadey
    happy to do a joint talk with you and michael. Thanks again for all the hard work contributing to this. It's really been very cool that you guys are so interested.
    Scott P. Jones
    @ScottPJones
    @swadey You wouldn’t by any chance be going to the NE Database Day, would you? (this Thursday)
    If you are around the Stata center, and you see an old fat guy, with a Mac laptop covered with Julia stickers (and Spock, Gandalf, etc), that’s me! :grinning:
    Ismael Venegas Castelló
    @SalchiPapa
    @/all I think I might not be able to go for this JuliaCon either as a speaker or as a listener :/ , but maybe next year, ...maybe at JuliaCon Mexico 201x? who knows!!! :P
    Ismael Venegas Castelló
    @SalchiPapa
    This message was deleted
    Ismael Venegas Castelló
    @SalchiPapa
    Oh damn how could I delete that! :O It was a long comment, anyway I will let you know when the acceptance for talk proposals is up :D
    S Wade
    @swadey
    @ScottPJones not on campus today unfortunately.
    @Ismael-VC that's too bad. It would be cool to talk about this but perhaps next year.
    Christian Weilbach
    @whilo
    hey, is this lisp still alive? looks cool to me :)
    Ismael Venegas Castelló
    @SalchiPapa
    @whilo it sure is! :D
    what do you think about it?
    S Wade
    @swadey
    yes. I haven't had a chance to get back to this but it's still a going concern...
    Christian Weilbach
    @whilo
    I am using mostly Clojure (did the scheme metacircular evaluator from SICP in ClojureC, some emacs lisp and have contributed a bit to Hy-lang). I haven't used Julia yet, as I don't like matlab because I think a general purpose Lisp is better to adapt to mathematics (Matrix slicing dsl etc.) than inventing a special syntax and runtime for special problems as is done by Julia (or R). But the JIT compiler of Julia is definitely more reasonable than CPython or the RVM which rely heavily on C-libs for performance. It would be nice to use this lisp directly from within Clojure for example, similar to using hy-lang in in Clojure for Python libs. This is only an idea though.
    I have to try it now, I guess :)
    Scott P. Jones
    @ScottPJones
    @whilo Is it just the Lisp syntax you prefer, or do you like Clojure because it has easy access to the rest of the JVM ecosystem? Many people prefer the Python/C/Java like syntax, as being more compact and readable than so many levels of parenthesis (although I did love Scheme’s minimalist beauty), and I wouldn’t consider that what Julia does as being just a “syntax and runtime for special problems”. I think the idea was more to have a syntax easier for mathematicians / scientists, while retaining all of the goodness of languages like Scheme (such as the metaprogramming). The other big thing about Julia is that it is dynamic, but with lots of static analysis, better ways of getting top performance than having to use another language (usually C or C++) for all the parts that need to be performant. @swadey’s nice work here let’s you use the syntax you prefer (Clojure-like) while retaining Julia’s nice features. (It’s funny, but at Jeff Bezanson’s PhD thesis [great reading, btw] defense, Gerry Sussman was one of the 3 on the panel, and afterwards he asked for something just like this! He’s also not a fan of too much syntax ;-) )
    Jorge Pérez
    @PerezHz

    Hi everyone! First of all I want to congratulate @swadey for his awesome job developing LispSyntax.jl! @ScottPJones kindly told me about this package, I have been experimenting with it (using also LispREPL.jl) a little bit and it works really nice!

    A couple of days I came across Gerry Sussman's Scheme and saw that he uses it in a really elegant way to describe classical mechanics, and became very interested in being able to bring his Scheme into Julia! I posted on JuliaLang/julia gitter and there @ScottPJones suggested me to contribute to this project

    So, how can I help you guys?

    Right now I'm a physics PhD student at UNAM, Mexico (actually met @Ismael-VC once in a Mexico City meetup last year :smile: when I was just starting to learn Julia ) , my main focus until now has been ODEs (using TaylorSeries.jl and hopefully will also contribute to ODE.jl as a GSoC student), I'm by no means an advanced Julian but it would be an honor to contribute a couple PRs!

    Ismael Venegas Castelló
    @SalchiPapa
    hi @/all !
    Ismael Venegas Castelló
    @SalchiPapa
    @perezhz it would be very useful if you just used LispSyntax, try to port snippets of other lisps, try to use the REPL exclusively for say 15 minutes, and I'm sure you'll still find lots of issues, that need to be ironed out, for example literal syntax for many things is not yet defined, and any other idea you have, welcome aboard! :D
    I was thinking that LispSyntax could use Lazy.list instead of array for it's data structure, which would differentiate between vectors (nd-arrays) and (linked) lists, the idea is that it should resemble Clojure.
    There is this other lang, called Hy, that is implemented in Python, but is inspired by clojure, but it is free to do it's own thing when reasonable, I mean it doesn't try to be a port of clojure for the python platform, it's just inspired by it. In the same way I think LispSyntax should stand on it's own, where possible, basically Julia is already a lisp, but a lisp with syntax, we just need to provide an alternative s-expression syntax, but leaving semantics the same.
    S Wade
    @swadey
    @perezhz welcome! This is early days, so any lisp code that you generate would be helpful so that we can test. Documentation, tests, etc. would also be a good.
    Scott P. Jones
    @ScottPJones
    @swadey, good to see you’re still alive! must be keeping busy!
    Jorge Pérez
    @PerezHz
    Thanks @Ismael-VC !! Well yesterday I was trying some basic examples I took from Gerry Sussman's books and they seemed to work fine!! Was also trying to
    Jorge Pérez
    @PerezHz
    Make some basic anonymous functions work and was able to run them without issues
    Those were basic examples but perhaps they could fit into the tutorial?
    Thanks @swadey so much! I will work on trying to reproduce Gerry Sussman's Scheme Mechanics
    Jorge Pérez
    @PerezHz
    Scheme Mechanics is a flavor of Scheme which Gerry uses along with Jack Wisdom for an advanced Classical Mechanics course at MIT
    Jorge Pérez
    @PerezHz
    I will work on that and add them as tests for LispSyntax.jl !
    Christian Weilbach
    @whilo
    @perezhz nice! structure and interpretation of classical mechanics is also on my reading list, ping back once you have done something.
    Christian Weilbach
    @whilo
    @ScottPJones I am a bit frustrated by the fragmentation of language systems. I had to learn so many languages which are almost the same, just to participate in a different (scientific community): Python, Matlab, R, Julia (currently playing around) for science, then Java, JavaScript, C++/C for application development, Erlang (currently playing around) and Go for distributed systems, Haskell, Ocaml etc (playing around atm.) for functional programming. I think it is insane when you see that you can implement any computational concept from the inside out in Lisps without baking it into syntax or making it mandatory for every user. The difference is too often starting with syntactic details, but also by baking runtime semantic details inside languages. I think Lisp still has important things to tell us about the universality of computing and a proper syntactic model for it. Separating the language from the runtime host is another important step by Clojure imo. I have started to play a bit with embedding modern lisps here, for example: https://github.com/whilo/hy-clj
    to participate in a different community I wanted to say
    Scott P. Jones
    @ScottPJones
    Yes, exactly. For me, that’s one of the attractions of Julia, it helps cut down on the number of languages I need to use / learn / become expert in. I had been learning Python, relearning C++ (with C++14), might have had to learn R, when I was told to investigate Julia a year ago (for the startup I’m at)
    Christian Weilbach
    @whilo
    For instance I would always implement a language as Lisp first (even staticly typed languages), so you can write macros and use sexps without needing to operate on full ASTs, and then put a less expressive syntax as convenience on top. Clojure, Hy and LispSyntax.jl are going the other way, which might help to communicate the concept of Lisp and is more realistic.
    Cool. But for me Julia is still a language for scientific computing, right? I won't use it to write a normal website any time soon, I think.
    Scott P. Jones
    @ScottPJones
    I’m hoping Julia will be good enough for any distributed and functional programming I want to do, also.
    Christian Weilbach
    @whilo
    Ok. What general purpose applications are build with Julia?
    Scott P. Jones
    @ScottPJones
    When people talk about Julia as a language for technical/scientific computing, it annoys me :grinning: , because although that is the initial focus, it’s a perfectly good language for general programming (much better, IMO, than many others).
    Christian Weilbach
    @whilo
    The problem I see is building a language and a runtime at the same time assumes that the language always has its host. Later missing runtime features need to be reimplemented anyway, when you e.g. try to target JavaScript runtimes.
    Scott P. Jones
    @ScottPJones
    I even asked Jeff about that, at his Ph.D. thesis defense - his response, if a language can handle the real tough issues of technical/scientific computing, then it should be able to easily handle issues of general programing.
    Christian Weilbach
    @whilo
    Hehe, well... Erlang for instance is build for specific realtime needs, but has poor performance for computation.
    Scott P. Jones
    @ScottPJones
    Yes, no one language to replace them all, I like the idea of Juliua as being the one language to rule them all, by binding easily to all
    Christian Weilbach
    @whilo
    Scientific computing is mostly about operating fast on numbers, while many systems benefit from a strong JIT-compiled runtime the most performancewise. Scientific computing needs are not common in industry, I would say.
    How does it easily bind?
    Scott P. Jones
    @ScottPJones
    To C (or other languages with a simple interface using the C ABI), it’s just a ccall.
    Then there are the PyCall and RCall packages.
    Christian Weilbach
    @whilo
    Ok, but you can do this for example on the JVM also (uhu ugly JVM :P)...
    Is the Julia runtime well-designed for concurrency?
    What I would love would be a runtime written in a subset of Lisp including efficient emitters for assembler (where code is also data), so you can manipulate your runtime from the inside while it is running. Calling C routines is always calling blackboxes...
    Scott P. Jones
    @ScottPJones
    I know many people are doing so, process based concurrency, real thread support is coming soon.
    Christian Weilbach
    @whilo
    Ok, because Clojure for example makes building multithreaded applications reasonable, but you need to build on immutability and safe memory primitives for that. Languages which encourage/allow imperative programming style are usually difficult to use for composition in multithreaded environments.
    Well, I hae to do some machine learning stuff, will stick around though. Nice talking to you :). Hopefully I can soon hack LispSyntax.jl a bit.