On Wed, 14 Oct 1998, Hans Aberg wrote:
> At 14:36 -0500 98/10/09, Jan Skibinski <[EMAIL PROTECTED]> wrote:
> > Could Haskell ever be used for serious scientific computing?
>
> I think you need to define "scientific computing": There is usually a
> tradeoff between runtime abstraction/generality and speed. Haskell provides
> some generality on the expense of speed.
I do not think it is possible to clearly define the phrase.
It may mean different things to different people.
Personally, I would extend it to "scientific/engineering
computing" to stress practical aspect of some computations.
A practicing engineer might be interested in job at hand
and not pretending doing any science whatsoever. Yet he/she
would need a reliable tool that could handle the size of
the problem (be it 10, 100, 1000, or million nodes or dimensions
) in reasonable time.
>
> Haskell could probably be made more usable by allowing classes be written
> in another language and used from Haskell or vice versa.
Good point!
Although not ideal but it has some practical connotation.
I think Ocaml is going this route, since many of its routines
are written in C, and their tools are excellent for interfacing
with C.
Jan