|
Boost : |
From: Paul Moore (gustav_at_[hidden])
Date: 2002-03-10 17:43:29
On Sun, 10 Mar 2002 17:31:03 -0500, Beman Dawes <bdawes_at_[hidden]> wrote:
>At 05:27 AM 3/8/2002, Moore, Paul wrote:
> >It's arguable that these sort of issues make rational<> inappropriate for
> >the standard library. And it's *definitely* arguable that
> >unlimited-precision integers are more appropriate, and should be
> >considered
> >first. (One slight issue - there's no candidate implementation of
> >unlimited-precision integers, yet!)
>
>That sounds reasonable. If you would like, I'll ask the committee to table
>rational pending unlimited-precision integers.
I'm not entirely convinced of the need to go that far. It's certainly
worth pointing out to the committee that rationals based on
limited-precision integers, although they are an "obvious" thing to many
users, suffer from subtle precision and rounding issues (assuming they
hadn't already noted the point).
If the committee feels that this makes rationals not ready for
standardisation in the absence of an unlimited-precision integer class,
I am happy with that. If they feel that there is still some value in
providing a canonical definition of rationals in any case, I'm happy to
go with that as well.
[In other words, as long as the committee is made aware of the issue,
that's enough for me - I don't have the experience to judge what way the
decision should go].
Paul.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk