On Sat, 17 Oct 1998, S.D.Mechveliani wrote:
> Only the meaning of the words `scientific computing' in programming
> has 90% changed since 1960-1970.
> Now it means mainly the *symbolic* (not approximate) computation that
> the scientists and engineers usually do on the paper.
I think that the terms `scientific computing' were being used in the sense
of `computing in order to acheive scientific goals' rather than any other
more technical meaning. My experience is again rather limited but I've
hung around one maths dept & two computer science departments and I'd say
that _at most_ 60% of people who have been computing have been doing
symbolic work only, at a conservative estimate. Most people these days do
some symbolic algebra but rarely at the sort of depth which requires
things beyond standard packages for Maple or Mathematica.
> The `number crunching engine' for the tasks like the large float matrix
> inversion constitutes maybe 5% of the scientific computing matter
> (still, it is important).
There's lots of numerically intensive algorithms out there which aren't
simple applications of classical methods.
> So testing the fitness for the scientific computing will rather mean
> to program, say, the polynomial factorization in Haskell, or maybe, the
> logical resolution method, and compare its performance to AXIOM, Maple,
> MuPAD ones.
This is definitely one area in which usefulness for scientific computing
should be tested, but it's by no means the only one. Maybe we should try
and get an informal poll (on this list) of the sorts of areas in which
haskell is currently, or might in the future, be used to discover other
areas of scientific computing to consider.
___cheers,_dave__________________________________________________________
email: [EMAIL PROTECTED] I _never_ buy items advertised by those
www.cs.bris.ac.uk/~tweed/pi.htm annoying cards that drop from magazines
work tel: (0117) 954-5253 or by bulk e-mail on principle.
