Hi, thx for this reply. Is there any overhead using this mutable?
I just thought I should point out that "Mutable" is not an haskell type. You can see in the Utils module that it is just a type synonim for IORef: http://icfpcontest.cse.ogi.edu/simulator/pfe.cgi?Utils#Mutable http://haskell.cs.yale.edu/ghc/docs/latest/html/base/Data.STRef.html I think that sometimes STRef might be a better choice...
Are you also using Templates ? With this mutable I can adapt a lot of software from Ocamel.
I think this John Hughes paper "Global Variables in Haskell" might be usefull to you: http://www.math.chalmers.se/~rjmh/Globals.ps
But the gnawing question remains: Shall it be possible to be almost as efficient (in native code) as Ocamel's code, I refer here to Doug Bagley's comparison of programming languages. Remarkably the fibonacci numbers test scores almost as well. But the code is not the same. Comparable code would have been to use the same code of Ocamel, to be more specific: how efficiently is recusivity implemented in Haskell. I cannot compare this on my Windows XP since I need MSVC6.0 on this machine which I don't have. On the other hand I have cygwin installed now. Unfortunately I can't make makefiles. Probably on the web I can find an explanation. In this way I can recompile Ocamel with Cygwin and compare the results a bit. There are also 5(?) failures of Haskell programms . Is there a flaw in these programms? P.S. does anyone know a good Haskell IDE for Windows XP?
I don't really use windows but I know there is support for the windows version of JCreator: http://www.students.cs.uu.nl/people/rjchaaft/JCreator/ You can always use Xemacs :) http://www.xemacs.org/Download/win32/ http://www.haskell.org/haskell-mode/ J.A.