[clean-list] Re: 1.Floating Point Support (clean-list digest)
Jerzy Karczmarczuk
karczma@info.unicaen.fr
Thu, 16 Nov 2000 10:12:33 +0000
Siegfried Gonzi wrote:
> I do not know what you really mean, but rounding errors are
> inavoidable (in everey language); that is my naive assumption.
>
> How can a programming language active control the rounding errors?
> A number is rounded up or down.
>
> But hasn't Clean incorporated the BigNumerical-Library for that
> issues?
Oh, my goodness...
There is a whole huge theory and practice around/behind that...
Perhaps you should have a look on the literature about the
floating point precision, what to do with "hidden bits", etc.
There are nice references here,
http://cch.loria.fr/documentation/IEEE754/index.html
for example the slides
"What can you learn about Floating-Point Arithmetic in One Hour ?"
by Kahan (who else...):
http://cch.loria.fr/documentation/IEEE754/wkahan/cs267fp.ps
[[of course the fact that I am giving french links is a bit
accidental, you find such stuff everywhere.]]
A MUST (one of favourite words of the new French generation) is
the article of David Goldberg: "What Every Computer Scientist
Should Know About Floating-Point Arithmetic", published in
Computing Surveys, March 1991, and available on-line
(for example here):
http://www.validgh.com/goldberg/paper.ps
You will bury very deeply your "naive assumptions"...
===
BigNums are not *FOR THAT* ! A bazooka to kill flies is OK, but
not very economic.
Jerzy Karczmarczuk
Caen, France