Subject: Maxima doesn't assume ommutativity? (bug?)
From: Daniel Lemire
Date: Sun, 03 Feb 2002 14:17:04 -0500
Good day,
The example I gave was too simple. Yes, a*b-b*a does simplifies to
zero... but see the slightly more challenging problem below...
Do you feel I should report this as a bug or am I being silly??? Please,
someone, do a sanity check on me! It seems to me that Maxima doesn't
assume commutativity!!!! (why???)
[lemire@romeo lemire]$ maxima
GCL (GNU Common Lisp) Version(2.5.0) jeu jan 31 23:03:22 EST 2002
Licensed under GNU Library General Public License
Contains Enhancements by W. Schelter
Maxima 5.6 jeu jan 31 23:03:13 EST 2002 (with enhancements by W. Schelter).
Licensed under the GNU Public License (see file COPYING)
(C1) expand((-A^3*a/64+a*A^3/64));
3 3
a A A a
(D1) ---- - ----
64 64
(C2) scsimp(D1);
3 3
a A A a
(D2) ---- - ----
64 64
(C3)
This type of bug crept up on my as I was doing some "real" research
using Maxima. I had a conjecture that such polynomial and such other
polynomial would simplify in such a way that... (skip details)... but I
spent (wasted) a day with very complicated formulas until a realized
that Maxima didn't assume commutativity!!! So that even though I
strongly suspected that things would greatly simplify, they apparently
did not!
Of course, here I give a rather simple examples, but it happened to me
several times over the last few weeks, involving various complicated
expressions. This one was the simplest I could find.
This problem seems to happen systematically, but is somewhat random at
the same time (that is, I can't tell from the expression whether this
bug will happen... but once it happens for a certain expression, if I
have this expression with other terms, the problem with happen! Hence,
it is systematical...)
I've recently upgraded my lisp. It didn't fix the problem for me.
I have
Linux 2.4.2-2 with a 686
Maxima built from cvs using latest lisp (from cvs). The problem happened
to me on the maxima RPM though.
I can't believe I would be the first one to notice this problem???
--
Daniel Lemire, Ph.D.
http://www.ondelette.com/