March 2001 - Haskell - Haskell.org

Re: constants and functions without arguments
by Andreas Leitner 30 Mar '01

30 Mar '01

Tom Pledger wrote: > > Andreas Leitner writes: > : > | Given a lazy pure functional language do we need to differntiate > | (in syntax) between constants and functions without agruments? And > | if we don't need to, does Haskell make a difference? > > Haskell always treats a declaration of the form > > foo = ... > > as a pattern binding, not as a function binding. Sections 4.4.3 and > 4.5 of the Haskell 98 report give more details. > > So, in that sense, every function in Haskell takes an argument. > There's nothing to prevent you from adding a dummy parameter to turn a > constant (i.e. a simple pattern binding) into a function. The only > reason I've heard for doing so, is to work around the monomorphism > restriction. Exactly thats' the reason for my question: Overcome ther monomorphism. I come from an Eiffel background and there we have the uniform access principle, which is a very important thing in that world. It means that you access an attribute just like a function with no parameters. This is actually a form of information hiding as you can change the underlying implementation (an attribute to a function and vice versa). The classic examlple for this is class POINT, which has an interface like that: -- POINT -- group 1 x: DOUBLE y: DOUBLE -- group 2 rho: DOUBLE abs: DOUBLE -- Now, you can choose to either implement group 1 as attributes, or group 2 and calculate the other one. With the uniform access principle, which says that accessing attributes and functions with no arguments happen in the same syntactic way, you can change the implementaion without notice. No user of this class needs to know. I am to new to FP to say whether this is equaly important in FP-languages, since the semantics are different (constants vs. attributes and immutable values vs. objects), but for now I am just currious wether it is possible theoretically. I mean couldn't one say that there are no constants, just functions with no arguments or the Void/Unit argument that return an expression. Since we have lazy evaluation, there won't be a problem at runtime, but would the type system allow such a thing? tia, Andreas

2 1

beginners silly question
by Luc taecsh 30 Mar '01

30 Mar '01

<intro> Ive just decided to start with haskell (after a decade of OO, and a few month of erlang (already impressive)) despite i love linux, id like to stay with wintel, to be able to use it at the office, and try to slot it in my daily job, slowly. (doing some modeling, should be useful) </intro> 1)are there a full package of haskell over there ? 2)ive installed hugs, SOE libs, haxml, frantk, tcl/tk, g/hood, hdoc, and jcreator what am i missing ? any incompatiblities +98 -98, and so on ? 3)if i get some good stuff done, id like to compile them , as small utilitaries. (best way to turn H in my daily job) my first try at ghc was not that fun (deppendencies on gcc, libs, and so on) as a alternative, ive tried under linux (m7.2),where ghc is installed by default, but even here , were i was expecting fluentness, its not cristal clear, hangs, missing bits i dont understand. no samples/ demo ? or is it a packaging issue ? <whining> 4) generally, are there any project to have some ready to use all-included package ? i understans some effort are under way on the great lib side, but are there anything avaialble now (its still farther than for me) for info, in the eiffel world, some guy (elj) once tried to put all the stuff together in a click and install package, and it really helped the gnu eiffel stuff starting in win world, (where things are not so simple , when unix is assumed with gnu stuffs...) (if some think answering DIY, just think its what effectivelly happen^s everytime, everywhere in the world someone goes beyong haskell.org front page... but all u haskell people are too smart for me to mention that.. ;-) </whining>

1 0

constants and functions without arguments
by Andreas Leitner 30 Mar '01

30 Mar '01

Hi, I hope this is the right forum to post my question to. Given a lazy pure functional language do we need to differntiate (in syntax) between constants and functions without agruments? And if we don't need to, does Haskell make a difference? In a language with eager evaluation (let's take SML) we must differentiat. Given we have a function that takes one argument of type int, is different than a function that takes unit->int. Because of that in SML, whenever I have a function (say "h") that takes an int, and I want to apply a function (without arguments), let's call it g) on it I need to add "()" at the end. h g() but for any constant i of type int I write: h i Now, this leads to my question: In a lazy language do I need to write those "()", or could the compiler tell the difference and always do the correct thing? many thanks in advance, Andreas

3 2

RE: communication between thread
by Simon Marlow 30 Mar '01

30 Mar '01

> I am using the library oncurrent of GHC. I'd like to know how two > different threads can communicate each other. > To solve the problem I thought in something like a global variable. If > there any method to avoid the use of that kind of variables > in haskell. You can use the various concurrent data structures in the Concurrent library. See: http://www.haskell.org/ghc/docs/latest/set/sec-concurrency-abstractions. html The simplest form of communication is through an MVar, which is a thread-safe mutable variable. Cheers, Simon

1 0

communication between thread
by Juan Ignacio García García 30 Mar '01

30 Mar '01

Hello, I am using the library oncurrent of GHC. I'd like to know how two different threads can communicate each other. To solve the problem I thought in something like a global variable. If there any method to avoid the use of that kind of variables in haskell. Thanks in advance J. I. García

1 0

RE: Parameterising Class Constraints in Existential Types
by Simon Peyton-Jones 30 Mar '01

30 Mar '01

It seems reasonable to me, and GHC is happy. Here's a working example: Simon class MyClass a b where foo :: a -> b -> Int data Special = forall b. (MyClass Int b)=> MkSpecial b data General a = forall b. (MyClass a b)=> MkGeneral b instance MyClass Int Bool where foo x False = -x foo x True = x xs :: [General Int] xs = [MkGeneral True, MkGeneral False] main = print [foo (3::Int) x | MkGeneral x <- xs] | -----Original Message----- | From: Ashley Yakeley [mailto:ashley@semantic.org] | Sent: 29 March 2001 10:47 | To: Haskell List | Subject: Parameterising Class Constraints in Existential Types | | | Is this valid (extended) Haskell? | | -- | class MyClass a b where | foo :: a -> b -> Int | | data Special = forall b. (MyClass Int b)=> MkSpecial b | data General a = forall b. (MyClass a b)=> MkGeneral b | -- | | Hugs complains about the 'General' line but has no problem with the | 'Special' line... | | -- | Ashley Yakeley, Seattle WA | | | _______________________________________________ | Haskell mailing list | Haskell(a)haskell.org | http://www.haskell.org/mailman/listinfo/haskell |

1 0

getEvent Doubts
by Andre W B Furtado 28 Mar '01

28 Mar '01

[Sorry about the previous email subject :) ] In Hugs Graphics Library, many procedures call the function getEvent with a Window as the parameter. But getEvent is defined as below: getEvent :: Events -> IO Event How does this work? I've tryied to do something similar and got the following error message (in GHC): Couldn't match `Events' against `Window' Expected type: Events Inferred type: Window In the first argument of `getEvent', namely `w' In a do statement: e <- getEvent w In time: my "Window" is defined as following data Window = MkWindow { win :: GLUT.Window, -- the real window events :: Events -- the event stream } And in Hugs Graphics Library the "Window" is defined as: data Window = MkWindow { wnd :: WND, -- the real window events :: Events, -- the event stream graphic :: IORef (Draw ()) -- the current graphic } Thanks, Andre

3 2

Difficulties compiling hmake / HaXml
by Lescher Christian 28 Mar '01

28 Mar '01

I want to use HaXml, and since hmake is required first, I'm currently trying to build hmake with GHC. Unfortunately "make" fails; I get the above errors message. What am I doing wrong here? I use GHC4.08.2 (InstallShield version) under WinNT4/Cygwin1.1.8. Christian make[1]: Entering directory `/test-hmake/hmake-2.02/src/interpreter' mkdir -p /test-hmake/hmake-2.02/targets/ix86-CYGWIN_NT-4.0/obj/interpreter /test-hmake/hmake-2.02/script/hmake -ghc HInteractive -d/test-hmake/hmake-2.02/t argets/ix86-CYGWIN_NT-4.0/obj/interpreter Fail: Can't find module Posix in . Or in standard libraries at /ghc/ghc-4.08.2/lib/imports/std /ghc/ghc-4.08.2/lib/imports/data /ghc/ghc-4.08.2/lib/imports/lang /ghc/ghc-4.08.2/lib/imports/posix /ghc/ghc-4.08.2/lib/imports/net /ghc/ghc-4.08.2/lib/imports/num /ghc/ghc-4.08.2/lib/imports/text /ghc/ghc-4.08.2/lib/imports/util /ghc/ghc-4.08.2/lib/imports/win32 /ghc/ghc-4.08.2/lib/imports/concurrent Asked for by: HmakeConfig.hs [Check settings of -I or -P flags?] Stop - hmake dependency error. make[1]: *** [/test-hmake/hmake-2.02/lib/ix86-CYGWIN_NT-4.0/HInteractive.exe] Er ror 1 make[1]: Leaving directory `/test-hmake/hmake-2.02/src/interpreter' make: *** [targets/ix86-CYGWIN_NT-4.0/hmake_g] Error 2

1 0

Duvidas
by Andre W B Furtado 28 Mar '01

28 Mar '01

In Hugs Graphics Library, many procedures call the function getEvent with a Window as the parameter. But getEvent is defined as below: getEvent :: Events -> IO Event How does this work? I've tryied to do something similar and got the following error message (in GHC): Couldn't match `Events' against `Window' Expected type: Events Inferred type: Window In the first argument of `getEvent', namely `w' In a do statement: e <- getEvent w In time: my "Window" is defined as following data Window = MkWindow { win :: GLUT.Window, -- the real window events :: Events -- the event stream } And in Hugs Graphics Library the "Window" is defined as: data Window = MkWindow { wnd :: WND, -- the real window events :: Events, -- the event stream graphic :: IORef (Draw ()) -- the current graphic } Thanks, Andre

1 0

Regarding open positions at IT-C
by rms 27 Mar '01

27 Mar '01

Regarding open positions at The IT-University of Copenhagen, Via the following URL you will find a detailed description of 13 positions as Associate or Assistant Professor currently open at The IT-University of Copenhagen. http://www.it-c.dk/pub_uk/db_363.html Kind Regards Personnel Department/ The IT-University of Copenhagen

1 0