- Glasgow-haskell-users

RE: Why is (monad) elegance so costly?
by Simon Peyton-Jones 03 Nov '01

03 Nov '01

| However, my experiments with a simplified lambda-calculus | example shows that (with GHC 5.00) the state monad is | dramatically less efficient than the simple identity monad: | | 4 TIMES SLOWER, and | 7 TIMES MORE MEMORY! | | Is this normal? Acceptable? Am I doing something wrong? I've just measured your program with GHC 5.02. The speed difference seems to be a factor of 2, not 4, though that is still not good. You can see what is going on if you give the flag -ddump-simpl to GHC, and then look for the function Main.eval. You'll see that eval has a shape like eval (Var x) = let ... in \env -> ... eval (Add u v) = let ... in \env -> ... This is bad, because eval is building a function closure for the \env, instead of taking its two arguments together as does simplEval. We'd prefer eval (Var x) env = let ... in ... eval (Add u v) env = let .. in ... Why doesn't GHC do this? Because doing so risks losing sharing. It's possible that you may say let v = eval term in do { v ; v ; v ; v } So you eval the term once, to get a state transformer, and then run that state transformer four times. If the 'env' arg is floated out, the work of the ".." in the above lets would be duplicated. Now *in this case* you aren't (ever) sharing a partial appliation of eval, and GHC could usefully spot this (but it doesn't at the moment). But if eval was exported, so GHC couldn't see all the applications of eval it could no longer do this trick. Maybe we could generate two versions. Another possibility would be to let the programmer give a pragma of some sort. Anyway I hope this helps explain where the costs come from. I'm interested in examples like this because it helps to know what analyses would be useful. Simon

2 1

Student Programming Projects
by Barbara Nostrand 03 Nov '01

03 Nov '01

Hi. Next Semester, I am supposed to teach a short course in Haskell. Can anyone recommend interesting programming projects which can be completed in about a month? I posted this message a while ago and got a few suggestions that sounded like they would maintain student interest. Alas, my disk drive crashed soon thereafter. So if you have given me suggestions before, please do so again. I was particularly intrigued by the suggestions put forth by someone who complained that someone else's suggested projects were too theoretical. We need to try to capture the students' imaginations here and many of them prefer concrete thought to abstract thought. Again. Thank you very much. Barbara Nostrand -- +----------------------------------+----------------------------------+ | Dr. Barbara Nostrand | Department of Computer Science | | Assistant Professor | SUNY College at Potsdam | | Computer Science and Engineering | Potsdam, New York | | (315) 267-2216 | 13676 | +----------------------------------+----------------------------------+ | Ignored domains: bestbiz.net, pop.net, hotmail.com, aibusiness.com | | vdi.net, usa.net, tpnet.pl, myremarq.com | | netscape.net, excite.com, bigfoot.com, public.com | | com.tw, eranet.net, yahoo.com, success.net | | mailcity.com, net.tw, twac.com, netcenter.com | | techie.com, msn.com | +---------------------------------------------------------------------+

2 1

Christmas Shopping and Theatre (optional) break in London from �38.75 pppn!
by Capital Worldwide 02 Nov '01

02 Nov '01

1 0

RE: First class modules (was Existential Typing)
by Simon Peyton-Jones 01 Nov '01

01 Nov '01

| artificial restriction. I think extending the system with | some form of | subtyping between records would be important for a usable | system. Your | paper suggested that you are planning to do this. Any | thoughts on current | ideas in this direction? Yes, we should really say something about this in the paper. A key design choice in our system is to use *nominal* not *structural* typing. The structural choice makes perfect sense, but it leads to complications we don't have (as well as benefits). So we don't have structural subtyping, and it would not be easy to add it. On the other hand, one could add nominal subtyping without too much difficulty: record A = { x,y :: Int } record B <: A = { z :: Int } We have not done that yet, but Mark thinks it would could be done without bad interaction with what we have already. Simon

2 1

RE: problem compiling
by Simon Marlow 01 Nov '01

01 Nov '01

> hello, I installed ok the rpm-packages on linux (Mandrake > 8.1), but when I > try to compile something the following message appears: > > f2.hs:1: > failed to load interface for `Prelude': > Bad interface file: ./Prelude.hi > does not exist > Action: openFile > Reason: No such file or directory > File: ./Prelude.hi > > If I go to the directory /usr/bin (where the binary file ghc > is) and I type > > ghc -c /home/user/file.hs > then the compilation is done without problems. > Does anyone know what I have to do for compiling without > problems in any > folder? > (The directory /usr/bin is my PATH). You probably have a file Prelude.hs in the current directory. The next release of GHC handles this case a little better, but to get going straight away you can just move Prelude.hs out of the way (and any other .hs files in the current directory which match the names of standard libraries). Cheers, Simon

1 0

RE: gtk+hs and libgmp
by Simon Marlow 01 Nov '01

01 Nov '01

> > That would be one solution, yes. On your system, if you > remove /usr/lib/libgmp.a (and /usr/lib/libgmp.so if you have > one), does it work? You only need these files if you intend > to build any software which depends on having GMP version 2, > which is unlikely since GMP 3 is source-compatible with GMP 2. > > Yes, that works. Still I would opt for the renaming to avoid this > problem in the future. It really depends on the setup of the system. > Another question: should packages like gtk+hs not add themselves to > package.conf, so that gtkhs-config is not needed? Or is this method > not yet proven. I've been discussing with Manuel how best to do this. It's not straightforward, because you still want to be able to do things like 'make install', and have several versions of gtk+hs on your system at once. Cheers, Simon

2 1

problem compiling
by Juan Ignacio García García 31 Oct '01

31 Oct '01

hello, I installed ok the rpm-packages on linux (Mandrake 8.1), but when I try to compile something the following message appears: f2.hs:1: failed to load interface for `Prelude': Bad interface file: ./Prelude.hi does not exist Action: openFile Reason: No such file or directory File: ./Prelude.hi If I go to the directory /usr/bin (where the binary file ghc is) and I type > ghc -c /home/user/file.hs then the compilation is done without problems. Does anyone know what I have to do for compiling without problems in any folder? (The directory /usr/bin is my PATH).

1 0

RE: gtk+hs and libgmp
by Gérard Milmeister 31 Oct '01

31 Oct '01

On Wed, 2001-10-31 at 15:49, Simon Marlow wrote: > > > Here is the output: > > 1. Main.out from a link that succeeds. > > 2. GtkHelloWorld.out, a sample program from gtk+hs, compiled with > > ghc -v -fglasgow-exts -lgtkHS GtkHelloWorld.hs `gtkhs-config > > --cflags` `gtkhs-config --libs` `gtk-config --cflags` > > `gtk-config --libs` > > > > In fact, you are right, there is a -L/usr/lib before -lgmp. > > Maybe the libgmp.a included > > with ghc should be renamed, for example libgmp_ghc.a > > That would be one solution, yes. On your system, if you remove /usr/lib/libgmp.a (and /usr/lib/libgmp.so if you have one), does it work? You only need these files if you intend to build any software which depends on having GMP version 2, which is unlikely since GMP 3 is source-compatible with GMP 2. Yes, that works. Still I would opt for the renaming to avoid this problem in the future. It really depends on the setup of the system. Another question: should packages like gtk+hs not add themselves to package.conf, so that gtkhs-config is not needed? Or is this method not yet proven. -- Gérard Milmeister Tannenrauchstr. 35 8038 Zürich gemi(a)bluewin.ch

1 0

RE: gtk+hs and libgmp
by Simon Marlow 31 Oct '01

31 Oct '01

> Here is the output: > 1. Main.out from a link that succeeds. > 2. GtkHelloWorld.out, a sample program from gtk+hs, compiled with > ghc -v -fglasgow-exts -lgtkHS GtkHelloWorld.hs `gtkhs-config > --cflags` `gtkhs-config --libs` `gtk-config --cflags` > `gtk-config --libs` > > In fact, you are right, there is a -L/usr/lib before -lgmp. > Maybe the libgmp.a included > with ghc should be renamed, for example libgmp_ghc.a That would be one solution, yes. On your system, if you remove /usr/lib/libgmp.a (and /usr/lib/libgmp.so if you have one), does it work? You only need these files if you intend to build any software which depends on having GMP version 2, which is unlikely since GMP 3 is source-compatible with GMP 2. Cheers, Simon

1 0

problem definind a function
by Juan Ignacio García García 31 Oct '01

31 Oct '01

hello, Does anybody know why it is not possible to define the folowing function in ghc? gggg= \x y -> (x+y - length [1..(gcd x y)] >= round ((x+1)/2)) ) The message obtained was this one: No instance for (RealFrac Int) arising from use of `round' at num-s-divs.hs:253 In the second argument of `(>=)', namely `round ((x + 1) / 2)' In a lambda abstraction: (((x + y) - (length [1 .. (gcd x y)])) >= (round ((x + 1) / 2))) Thanks in advance, J. I. García

2 1