From: Phil Richards (news_at_[hidden])
Date: 2004-01-09 07:37:28
On Fri, 09 Jan 2004 11:51:59 +0000, Andy Little wrote:
> Yes you could probably get most things done with rationals. OTOH I can now
> get most(all) done with integer powers.
Well, for most of the last year I've been using a (much simpler version
of what I now have) and it only used integer powers. Most of the time
it has been OK - though there have been a few equations where temporaries
end up involving the square-root of some odd-powered dimension.
> Frankly I have got the message long ago that my class is not good enough
> for boost and never will be. There is always some thing that can be
> found that it doesnt do. As Paul Bristow pointed out any review of my
> lib is a foregone conclusion AFAIK.
That may or may not be the case - the way I view these discussions is to
try and get something in boost that does the job I need. I don't actually
care very much whether it is written by Matthias, you, me, or somebody
else :-) However, unless approaches are tried and publicised there never
will be a library that does this job in boost - if what you have done
works for you, then at least it is worth it for you - but it is also worth
it to make other people *think* about the possibilities.
I really would like to get my library posted - then people can rip that
apart and I will feel less guilty about making negative comments about
other people's attempts. And hopefully it would give a few more options
for something "in the middle".
-- change name before "@" to "phil" for email
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk