James Amundson <amundson@users.sourceforge.net> writes:
> Good. I was hoping it would be that simple. I think the maxima-cmucl
> binary distribution will be more attractive if it doesn't depend on an
> external cmucl rpm.
>
> > This would be pretty nice. I update my cmucl quite often and that break
> > maxima since I don't update that as often.
>
> Yes. Only recently have I appreciated how inconvenient it is to have
> your installed maxima distribution depend on your installed lisp
> distribution.
Nevertheless, the ideal world would be one where Maxima would run on every
ANSI CL implementation, because this would give the possibility to load it
as an additional module if desired.
In fact, I'm thinking about using Maxima's polynomial manipulation for
computing finite element shape functions and quadrature rules in my toolbox
"Femlisp" [1]. I would like this to be an optional choice, mainly because
1. the overhead of Maxima is not really negligible,
2. portability.
(On the other hand, Maxima will probably become necessary in Femlisp for
very-high-order finite elements because of multiprecision, special
polynomials, and other goodies.)
Nicolas.
[1] See www.femlisp.org. Femlisp is only in a testing phase right now and
works only with CMUCL.