Re: [Haskell-cafe] Quick, somebody do something!
On Thu, Jan 14, 2010 at 11:21 PM, Don Stewart d...@galois.com wrote: While the month-old Go language makes the top 15? Methods considered unsound. I fully agree. But anyway, I don't think people either already in the haskell world or about to enter it will find this relevant. -- Alp Mestan http://alpmestan.wordpress.com/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] ANN: hnn-0.1, a haskell neural network library
Hi, I just released the first version of my Haskell Neural Network library. It provides the very minimal features anyone would need in a neural network library. It has yet to be completed (regarding the features) and there are some ways for opitmizations. Though, I'd be very glad to hear from Haskellers about this library, get feedback, etc. Some links to get started with HNN : http://www.haskell.org/haskellwiki/HNN http://hackage.haskell.org/package/hnn -- hackage page http://mestan.fr/haskell/hnn/ -- online documentation http://alpmestan.wordpress.com/2009/12/23/hnn-0-1-has-been-released/ Please don't hesitate to try it and play with it and give as much feedback as you want ! Haskelly yours. -- Alp Mestan http://alpmestan.wordpress.com/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] dsl and gui toolkit
Note that the Qt library supports CSS, and it's pretty fun and easy to use. On Wed, Oct 7, 2009 at 5:05 AM, John A. De Goes j...@n-brain.net wrote: Then change to early generation language. Point being CSS has plenty of pioneering flaws. Regards, John A. De Goes N-Brain, Inc. The Evolution of Collaboration http://www.n-brain.net|877-376-2724 x 101 On Oct 6, 2009, at 7:52 AM, Richard O'Keefe wrote: On Oct 7, 2009, at 5:47 AM, John A. De Goes wrote: CSS is a good start by it's beset by all the problems of a 1st generation presentation language, and is not particularly machine-friendly. Considering that CSS is _at least_ a 2nd generation language (it was preceded by DSSSL), that's rather funny. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe -- Alp Mestan http://alpmestan.wordpress.com/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Re: dsl and gui toolkit
It'd be great if you could at least paste some sample lines of usage of your library. On 10/5/09, Gregor Burger burger.gre...@gmail.com wrote: hi there, i can't resist to post grun here! although its python based the ideas behind are great! maybe some excellent haskell hackers can borrow some ideas from it. http://pypi.python.org/pypi/grun/ gregor 2009/10/5 Günther Schmidt gue.schm...@web.de lemme have it please! Günther Am 05.10.2009, 12:42 Uhr, schrieb Andrew U. Frank fr...@geoinfo.tuwien.ac.at: writing a gui is a mess (independent of wx or gtk) - too much detail is shown and not enough abstraction is done. haskell can help. i have written an experimental way of producing the GUI automatically with a description of the semantics of the types and operations involved (a la ontology, evnetually comparable what protege produces). the input is a descriptionof the entity ypes, the fields used, the functional dependencies between the fiels, plus the operations used. the division in screens and their layout. the rest ist automatic. the result is a GUI (with preferably gtk but i had also a wx version running). the ideas were inspired by eliot conal's work and wxgeneric, which seemed for administrative applications either too restricted or to specific. if somebody wants to try it out for his application, please write fr...@geoinfo.tuwien.ac.at (there is not much documentation and the code is not yet completely clean - testing by somebody else would be very valuable!) andrew ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe -- Alp Mestan http://alpmestan.wordpress.com/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] Re: Strong duck typing / structural subtyping / type class aliases / ??? in Haskell
I had never seen this work, it's just awesome ! And it only needs few Haskell extensions. Is this work deeply documented somewhere except in research papers ? If not, it could be worth doing, IMO. On Tue, Sep 29, 2009 at 9:37 AM, o...@okmij.org wrote: Alp Mestan wrote: Indeed, OCaml has stuctural polymorphism, it's a wonderful feature. *# let f myobj = myobj#foo Hi !;; val f : foo : string - 'a; .. - 'a = fun* And Haskell has that too: -- This is how we define labels. data Field1 deriving Typeable; field1 = proxy::Proxy Field1 -- This is how record selection looks like. foo f = f # field1 The inferred type of foo is *OCamlTutorial :t foo foo :: (HasField (Proxy Field1) r v) = r - v It doesn't seem too different from the OCaml's type; the type variable r acts as a row type. The quoted example is the first from many others described in http://darcs.haskell.org/OOHaskell/OCamlTutorial.hs The file quotes at length OCaml's Object tutorial and then demonstrates how the OCaml code can be written in Haskell. When it comes to objects, structural subtyping, width and depth subtyping, etc., Haskell does not seem to miss match compared to OCaml. In contrast, Haskell has a few advantages when it comes to coercions (one does not have to specify the type to coerce to, as Haskell can figure that out). The other files in that directory give many more example of encoding C++, Eiffel, OCaml patterns. -- Alp Mestan http://blog.mestan.fr/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Strong duck typing / structural subtyping / type class aliases / ??? in Haskell
On Fri, Sep 25, 2009 at 8:14 PM, Job Vranish jvran...@gmail.com wrote: Supposedly OCaml has an OO feature that does this but I haven't tried it out. Indeed, OCaml has stuctural polymorphism, it's a wonderful feature. *# let f myobj = myobj#foo Hi !;; val f : foo : string - 'a; .. - 'a = fun* IIRC, there has been work on Template Haskell for structural polymorphism. -- Alp Mestan http://blog.mestan.fr/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Parallel graphics
This is just awesome indeed. You should create a haskell wiki page about that, so that beginners could see Haskell can do that (TM) (yeah, some beginners doubt of it). On Thu, Sep 24, 2009 at 1:02 PM, Olex P hoknam...@gmail.com wrote: Awesome! On Thu, Sep 24, 2009 at 7:47 AM, Peter Verswyvelen bugf...@gmail.comwrote: This is seriously cool stuff!!! Maybe it's time to start a Haskell Demo Scene :-) (what's a demo scene? See http://en.wikipedia.org/wiki/Demoscene ) PS: Note that Conal Elliott also was generating GPU code using Haskell with Vertigo back in 2004: http://conal.net/papers/Vertigo/ On Thu, Sep 24, 2009 at 5:32 AM, Claude Heiland-Allen claudiusmaxi...@goto10.org wrote: Andrew Coppin wrote: (OK, well the *best* way is to use the GPU. But AFAIK that's still a theoretical research project, so we'll leave that for now.) Works for me :-) http://claudiusmaximus.goto10.org/cm/2009-09-24_fl4m6e_in_haskell.html There doesn't need to be a sound theoretical foundation for everything, sometimes sufficient ugly hackery will make pretty pretty pictures... Claude -- http://claudiusmaximus.goto10.org ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe -- Alp Mestan http://blog.mestan.fr/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] adding state in GUIs (qtHaskell)
On Fri, Sep 11, 2009 at 5:55 PM, Jeremy O'Donoghue jeremy.odonog...@gmail.com wrote: snip I don't have anything as neat to show you as Duncan's suggetion (I'd also be interested to see a cleaner way to do it - this sort of code always grates a little with me, although all of the major Haskell GUI bindings seem to need a similar programming style. However, at the most basic 'trying it out' level, I suspect that something very like this will work just as well for qtHaskell as it does for wxHaskell. Regards Jeremy Very interesting code. However, I'd be very curious to see if qthaskell handles .ui files. And how it does. With C++, thanks to the 'uic' command line tool, we generate a class from the .ui file, and then just have to store an instance of it in our window/dialog/widget/whatever. This class has a setupUI member function, taking a QWidget*/QDialog*/QMainWindow*/whatever, which initializes all the ui components and put them on our widget just like we asked it to do in the designer. Actually, I'm wondering how the trick could be done (and if it is already done ?) in Haskell without letting too much things generated and compiled at the C++ level with some FFI magic. -- Alp Mestan http://blog.mestan.fr/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Parsec for C or C++
There is a C++ parser in C++, it may be of help : http://42ndart.org/scalpel/ It's a quite advanced WIP. On Fri, Jul 17, 2009 at 3:58 AM, Sterling Clover s.clo...@gmail.com wrote: A parser for JavaScript (admittedly a much simpler beast) is part of Brown's WebBits: http://hackage.haskell.org/packages/archive/WebBits/0.15/doc/html/ BrownPLT-JavaScript-Parser.html Cheers, Sterl. On Jul 16, 2009, at 1:40 PM, Roy Lowrance wrote: Turns out that Language.C uses alex and happy. I'm looking to use Parsec. So back to the original question: Does anyone know of a C or java parser written using Parsec? - Roy On Thu, Jul 16, 2009 at 12:43 PM, Roy Lowranceroy.lowra...@gmail.com wrote: Thanks Rick. A perfect tip! - Roy On Thu, Jul 16, 2009 at 12:29 PM, Rick Rrick.richard...@gmail.com wrote: There is language.c http://www.sivity.net/projects/language.c/ http://hackage.haskell.org/package/language-c From a parsing standpoint, C++ is a massive departure from C. Good luck though. On Thu, Jul 16, 2009 at 12:25 PM, Roy Lowrance roy.lowra...@gmail.com wrote: I am working on a research language that is a variant of C. I'd like to use Parsec as the parser. Is there an existing Parsec parser for C or C++ (or Java) that could serve as a starting point? Thanks, Roy ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe -- The greatest obstacle to discovering the shape of the earth, the continents, and the oceans was not ignorance but the illusion of knowledge. - Daniel J. Boorstin -- Roy Lowrance home: 212 674 9777 mobile: 347 255 2544 -- Roy Lowrance home: 212 674 9777 mobile: 347 255 2544 ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe -- Alp Mestan http://blog.mestan.fr/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] Backpropagation implementation for a neural net library
Dear List, I'm working with a friend of mine on a Neural Net library in Haskell. There are 3 files : neuron.hs, layer.hs and net.hs. neuron.hs defines the Neuron data type and many utility functions, all of which have been tested and work well. layer.hs defines layer-level functions (computing the output of a whole layer of neurons, etc). Tested and working. net.hs defines net-level functions (computing the output of a whole neural net) and the famous -- but annoying -- back-propagation algorithm. You can find them there : http://mestan.fr/haskell/nn/html/ The problem is that here when I ask for final_net or test_output (anything after the train call, in net.hs), it seems to loop and loop around, as if it never gets the error under 0.1. So I was just wondering if there was one or more Neural Nets and Haskell wizard in there to check the back-propagation implementation, given in net.hs, that seems to be wrong. Thanks a lot ! -- Alp Mestan ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Backpropagation implementation for a neural net library
On Mon, Jun 15, 2009 at 5:00 PM, Trin Trin trin...@gmail.com wrote: Hi Alp, - even with correctly programmed back-propagation, it is usually hard to make the net converge. Yeah, I know, that's why we're training it until the quadratic error goes under 0.1. - usually you initialize neuron weights with somewhat random values, when working with back-propagation. Yeah, that'll be done too, once the algorithm will be ready. I'll provide fancy and easy functions to create a neural net just giving the numbers of layers and their sizes. - do some debug prints of the net error while training to see how it is going Good idea, yeah. - xor function cannot be trained with a single layer neural net !!! That's why there are 2 layers there, one hidden and the output one. I consider the inputs as ... inputs, not as a first layer of the NN. Thanls for your time. If you have any clues when reading the code, don't hesitate of course. -- Alp Mestan http://blog.mestan.fr/ http://alp.developpez.com/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Generics Versus Parametric Polymorphism
Parametric polymorphism is kinda boring in Haskell -- since it's been there since the beginning. We tend to reserve the term generics for higher order, and fancier, polymorphism. These kinds of things: http://hackage.haskell.org/packages/archive/pkg-list.html#cat:generics -- Don I think he was refering to Java Generics, this kind of things. Java's generics are much less powerful than parametric polymorphism in Haskell. -- Alp Mestan In charge of the C++ section on Developpez.com. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
