Actually, copying down the rules for <interval> [op]<interval> from
some text or reference work is not hard. In fact, it should be
possible to copy it out of a program or just use existing lisp
code. There are some subtleties regarding rounding at
endpoints.
The difficult part is integrating into maxima the rules that
violate the conventions already in maxima, such as
x-x --> 0 for all x.
This is false for intervals.
RJF
----- Original Message -----
From: Robert Dodier <robert.dodier at gmail.com>
Date: Wednesday, May 3, 2006 7:44 am
Subject: Re: [Maxima] big float arithmetic & mgrp, also Interval Arithmetic project
> On 5/3/06, Raymond Toy <raymond.toy at ericsson.com> wrote:
> > >>>>> "Robert" == Robert Dodier <robert.dodier at gmail.com> writes:
> >
> > Robert> With appropriate macros (to expand that into
> matchdeclare> Robert> and tellsimp or tellsimpafter) and
> function definitions,
> > Robert> this yields stuff like
> >
> > Robert> (interval (a, b) + interval (5, 7))^3;
> > Robert> => interval ((a + 5)^3, (b + 7)^3)
> >
> > What if, instead of a cube, you had a square? Then the resulting
> > interval would depend on the actual values of a and b and their
> > relationship to 5 and 7.
>
> Sure. But I'm assuming that the details were worked out
> long ago, and what makes this interesting to do in Maxima
> is not the difficulty of implementing the formulas.
> It's clear, I think, where to direct further effort -- namely
> the functions add_interval_noninterval, etc. Maybe
> someone wants to do that.
>
> Robert
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>