can a lazy language give fast code?Fair answer.

D. Tweed tweed@compsci.bristol.ac.uk
Wed, 31 Jul 2002 13:59:04 +0100 (BST)


On Wed, 31 Jul 2002, Scott J. wrote:

> I don't think I have got a fair answer to my questions regarding these
> (silly?) benchmarks. I cannot write the programs with the unboxed "things",
> but I have both the Ocaml compiler and the latest Glasgow compiler installed
> on my windows XP machine. So, if someone sends the programs I'll type it in
> and let you know these results. I don't want to be impolite : the fact that
> I am on this list proves that I am seriously interested in the elegance of
> Haskell. But I am searching a language to program in it: I think e.g. to a
> front end of the Lout typesetting program. Also I have the impression that
> such fancy things as HOpenGL are not for windows because of the GTK
> bindings. It seems that I have to move to a Linux OS.

My messages were more addressing the point which came up about what the
aims of benchmarking `ought to be' rather than addressing the question. It
seems to me most of the most responses are to the question `could a lazy
language compiler be written to give fast code' and you're looking at the
question `are there settings and programming idioms for a current compiler
that give fast code'. I'll leave it up to the much better qualified
various experts with the various compilers to give detailled advice.

___cheers,_dave_________________________________________________________
www.cs.bris.ac.uk/~tweed/  |  `It's no good going home to practise
email:tweed@cs.bris.ac.uk  |   a Special Outdoor Song which Has To Be
work tel:(0117) 954-5250   |   Sung In The Snow' -- Winnie the Pooh