[clean-list] Floating Point Support

[RT Happe]

Some recent posts touched numerical issues.  I'd like to take this
occasion to point out that Clean's floating point support leaves
quite a lot of room to improvement (as of R1.3).
    Most hardware conforms, more or less well, to the IEEE 754 floating
point standard;  most programming languages don't support standard features
(precisions, rounding modes, exceptions) adequately;  Clean doesn't
either.  These features are essential tools for the (civilised) floating
point programmer who wants to control rounding error accumulation,
check wether a computation exceeds the capacity of the platform,
use the hardware economically.
    I see or guess that support of IEEE 754 within the pure functional
paradigm is neither easily done nor academically rewarding.  If 
nonetheless one or the other maker of a general purpose programming
language designed for the development of real world applications wants to
(re)consider this issue, he may check out the lecture notes of (ACM Turing
award winner) W. Kahan

    http://http.cs.berkeley.edu/~wkahan/ieee754status/ieee754.ps

(including "ruminations on programming languages") and other references
listed at

    http://cch.loria.fr/documentation/IEEE754/
    http://babbage.cs.qc.edu/courses/cs341/IEEE-754references.html

rthappe