I've found what I think is a bug in vect.mac. Here's an example:
load (vect);
declare (n, nonscalar)$
declare (r, nonscalar)$
declare (a, scalar)$
-a*r~n;
The current result is
- a (- n) ~ r
In vect.mac, there is this declaration:
TELLSIMP(ETRUE~LESSP, -LESSP~ETRUE) $
If I change this to:
TELLSIMP(ETRUE~LESSP, -(LESSP~ETRUE)) $
the above expression reduces to:
a n ~ r
which is better.
However, I'm not very familiar with TELLSIMP, so before I submit this
as a bug and patch, I thought I'd ask for comments.
--
-- Stephe