- Haskell-Cafe - Haskell.org

a question concerning type constructors
by Eric 05 May '07

05 May '07

Hi all, In Haskell, is it possible to declare a type constructor with a variable number of type variables e.g. data Tuple * .... allowing the following declarations: t: Tuple u: Tuple Bool v: Tuple Bool Int w: Tuple Bool Int Char ? E.

3 2

Detect Either Windows or Linux environment
by SevenThunders 05 May '07

05 May '07

Is there a simple way to detect what operating system a given Haskell program is running under? It would help to make some programs that might have to interact with the operating system more portable. I haven't been able to quite figure out a simple way to do this. -- View this message in context: http://www.nabble.com/Detect-Either-Windows-or-Linux-environment-tf3695277.… Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.

5 5

Re: Monad definition question
by oleg＠pobox.com 05 May '07

05 May '07

Ilya Tsindlekht wrote > > It may be useful to relate to imperative programming: > > m1 >>= (\x -> m2) > > is > > let x = m1 in m2 > The analogy is not always straight-forward - try the list monad. This equivalence holds even for the List Monad. Here is an example of non-determinism: http://caml.inria.fr/pub/ml-archives/caml-list/2007/02/1d4df471019050b89e20… and here's an excerpt from that message, showing lets (aka bind): let numbers = List.map (fun n -> (fun () -> n)) [1;2;3;4;5];; let pyth () = let (v1,v2,v3) = let i = amb numbers in let j = amb numbers in let k = amb numbers in if i*i + j*j = k*k then (i,j,k) else failwith "too bad" in Printf.printf "got the result (%d,%d,%d)\n" v1 v2 v3;; Here, 'amb' is like 'msum' and any error is mzero. If we replace 'let' with 'do', and '=' with '<-' in some places, it looks almost like Haskell. Although the scheduler (the `run' function of the `monad') described in the message accumulates all possible worlds, it returns as soon as one `thread' made it successfully to the end. It is trivial to get the scheduler to run the remaining `treads' searching for more answers (by threads I certainly don't mean OS threads). The user code above stays as it is. The implementation of amb has been extended to annotate choices (and so possible worlds) with probabilities, which propagate in due course. The result not merely lists the answers but also their computed probabilities. > > Filinski then showed that the latter seemingly missing aspect > > indeed only appears to be missing. > Would this require some kind of macros doing extensive pre-processing > of the code? Hopefully the code above showed that no macros and no preprocessing required. In that code, addition, multiplication, if-then-else and all other OCaml operations remain as they have always been, with no modifications or redefinitions whatsoever. Delimited continuations suffice for everything, as Filinski proved in his paper.

1 0

Re: Monad definition question
by oleg＠pobox.com 05 May '07

05 May '07

Ilya Tsindlekht wrote: > Does the definition of monad silently assume that if f and f' are equal > in the sense that they return the same value for any argument o correct > type then m >>= f = m >>= f' Of course NOT! Here's an example, in a State monad f x = put True f' x = put False Clearly, _by the definition above_, f and f' are the same -- for any argument of correct type, they return the same value, namely, (). However, if we perform the observation execState (return 'a' >>= f) True execState (return 'a' >>= f') True we quite clearly see the difference. Robin Green wrote: > How could it be otherwise? How are you going to distinguish between f > and f' if they are indistinguishable functions, in Haskell? Because f and f' are NOT referentially transparent functions. They are NOT pure functions, their application may have _an effect_. And comparing effectful computations is quite difficult. It's possible: I believe (bi)simulation is the best approach; there are other approaches. It may be useful to relate to imperative programming: m1 >>= (\x -> m2) is let x = m1 in m2 Indeed, monadic 'bind' is *exactly* equivalent to 'let' of impure eager languages such as ML. The first monadic law return x >>= f === f x is trivial because in eager languages, any value (which is an effectful-free computation) is by default injected into the world of possibly effectful expressions: Any value is an expression. The second law m >>= (\x -> return x) === m becomes let x = e in x === e and the third law (m1 >>= (\x -> m2)) >>= (\y -> m3) === m1 >>= (\x -> m2 >>= \y -> m3) provided x is not free in m3 becomes let y = (let x = m1 in m2) in m3 === let x = m1 in let y = m2 in m3 So, `bind' is `let' and monadic programming is equivalent to programming in the A-normal form. That is indeed all there is to monads. Here's the paragraph from the first page of Filinski's `Representing Monads' (POPL94) "It is somewhat remarkable that monads have had no comparable impact on ``impure'' functional programming. Perhaps the main reason is that -- as clearly observed by Moggi, but perhaps not as widely appreciated in the ``purely functional'' community -- the monadic framework is already built into the semantic core of eager functional languages with effects, and need not be expressed explicitly. ``Impure'' constructs, both linguistic (e.g., updatable state, exceptions, or first-class continuations) and external to the language (I/O, OS interface, etc.), all obey a monadic discipline. The only aspect that would seem missing is the ability for programmers to use their own, application-specific monadic abstractions -- such as nondeterminism or parsers [31] -- with the same ease and naturality as built-in effects." Filinski then showed that the latter seemingly missing aspect indeed only appears to be missing. It is important to understand that once we come to monads, we lost referential transparency. Monadic code is more difficult to reason about -- as any imperative code. One often sees the slogan that Haskell is the best imperative language. And with monad, it is. One often forgets that 'best' here has the down-side. Haskell amplifies both advantages _and_ disadvantages of imperative programming. At least imperative programmers don't have to think about placing seq at the right place to make sure a file is read from before it is closed, and don't have to think about unsafeInterleaveIO. It seems the latter has become so indispensable that it is recommended to Haskell novices without a second thought. One may wonder if functional programming still matters.

2 1

Re: [Haskell-cafe] Debuggingy
by Monang Setyawan 04 May '07

04 May '07

On 5/5/07, Stefan O'Rear <stefanor(a)cox.net> wrote: > On Sat, May 05, 2007 at 11:36:16AM +0700, Monang Setyawan wrote: > > Hi, I'm a beginner Haskell user. > > > > Is there any way to trace/debug the function application in GHC? > > Absolutely! > > stefan@stefans:/tmp$ ghci X.hs > ___ ___ _ > / _ \ /\ /\/ __(_) > / /_\// /_/ / / | | GHC Interactive, version 6.7.20070502, for Haskell 98. > / /_\\/ __ / /___| | http://www.haskell.org/ghc/ > \____/\/ /_/\____/|_| Type :? for help. > > Loading package base ... linking ... done. > [1 of 1] Compiling Main ( X.hs, interpreted ) > Ok, modules loaded: Main. > *Main> :break fac Great!! Thanks, it really helps. I should update my GHC to the newest version (I use the old 6.4.2 with no break command) Is there any editor/IDE supporting this break command? It should be cooler if we can debug functions just by placing mark in the line. > > Stefan > -- Demi masa.. -- Demi masa..

1 0

Debugging
by Monang Setyawan 04 May '07

04 May '07

Hi, I'm a beginner Haskell user. Is there any way to trace/debug the function application in GHC? -- Demi masa..

2 1

Re: [Haskell-cafe] [IO Int] -> IO [Int]
by Daniel Fischer 04 May '07

04 May '07

> -----Ursprüngliche Nachricht----- > Von: "Rich Neswold" <rich.neswold(a)gmail.com> > Gesendet: 04.05.07 19:31:53 > An: Phlex <Phlex(a)telenet.be> > CC: haskell-cafe(a)haskell.org > Betreff: Re: [Haskell-cafe] [IO Int] -> IO [Int] On 5/4/07, Phlex <Phlex(a)telenet.be> wrote: > Hello all, > > I'm trying to learn haskell, so here's is my first newbie question. > I hope this list is appropriate for such help requests. > > I'm trying to write a function with the signature [IO Int] -> IO [Int] > > > Control.Monad has a function (called "sequence") that does this for you. In fact, "sequence" is a more generic solution since its signature is ( > Monad m => [m a] -> m [a]). Unfortunately, this won't work for infinite lists either, for those you'd need an 'unsafe' action, namely 'unsafeInterleaveIO', like import System.IO.Unsafe conv :: [IO a] -> IO [a] conv [] = return [] conv (x:xs) = do a <- x as <- unsafeInterleaveIO (conv xs) return (a:as) *Test> :set -fno-print-bind-result *Test> xs <- conv $ map return [1 .. ] *Test> take 20 xs [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20] > > As a newbie, I found it educational to peruse the various modules in "http://haskell.org/ghc/docs/latest/html/libraries/ > ". Just start looking at modules that sound interesting and see what has already been defined. Some modules at first may be too advanced, but if you go back to them in a few days (weeks?), they'll start making more sense, too. > Very good advice, I think, and it's also very instructive to read the source code, you can learn a lot from that. > > -- > Rich HTH, Daniel

2 1

Poor first impression
by Fernando Cassia 04 May '07

04 May '07

I admit in shame never having heard about Haskell before. I know about PHP, Python, IBM' s REXX, TCL, TCL/TK, perl... but Haskell, never. So, here's how I landed in Haskell-land: I was looking for a simple ncurses-based text mode mp3 player with some sort of basic GUI and found "HMP3" written in, you guessed it, Haskell. So I follow the directions and download the huge 30MB+ ghc-6.6.1-i386-unknown-linux.tar.bz2. Bunzip2 it. tar xvf it. ./configure and "make install". So far, so good. and I get the following message, supposedly telling me that the haskell compiler was installed OK... ======================================================================= Installation of ghc-6.6.1 was successful. To use, add /usr/local/bin to your PATH. For documentation, see /usr/local/share/ghc-6.6.1/html/index.html ======================================================================= (/usr/local/bin is already in my path) So I decide to call the ghc compiler with no arguments to see if it was indeed installed, and I get this: [root@localhost ghc-6.6.1]# ghc /usr/local/lib/ghc-6.6.1/ghc-6.6.1: error while loading shared libraries: libreadline.so.4: cannot open shared object file: No such file or directory # So, I conclude that Haskell is not ready for prime time, if it cannot install itself correclty including shared libs in a standard Fedora Core 6 system. Goodbye Haskell, I just wanted to compile a MP3 player, and perhaps if the compiler installed OK with no issues, I'd have taken a look at the language. But as of right now, I don't have time to waste with broken compiler installers. Byebye FC

19 29

outputing .hs file without comments
by José Miguel Vilaça 04 May '07

04 May '07

Hi all, Is there a simple tool or command to remove all comments from a Haskell file, i.e. something that outputs the input file but without any comments on it? Best Miguel Vilaça

4 3

Hugs dropped Text.Regex?!
by John Goerzen 04 May '07

04 May '07

I learned today, as I tried to run my ftphs unit tests under Hugs, that Hugs dropped Text.Regex between 200503.08 and 200609.21. The Hugs download page claims that they did this because the new GHC implementation of Text.Regex is not compatible with Hugs. But why can't Hugs keep the old implementation? This is going to render Hugs just about useless for me. I like regexps. -- John

2 1