I think that Schelter contributed some code
that runs only in GCL or is linked to GCL through
foreign function interface that makes some procedures
run far faster than expected. If you are not using
these procedures, the difference would not be visible.
I do not know where these are used but I suspect
polynomial factorization.
I have explored with Franz Inc the possibility of
their providing a base of Allegro CL for Maxima
development.
Advantage: supported ANSI CL on Windows, Mac OS-X, many
different UNIX and linux systems, many extras like
graphics, web support, etc. (see franz.com).
Disadvantage: it does not come with source code.
Disadvantage: there will be presumably SOME restriction
on this so that they are not giving away their main
source of revenue free. Yet this may not be much of
a restriction since Franz already gives away a rather complete
lisp free (minus some compiler features and maybe a
maximum memory limit, plus.. one must re-register periodically...).
The free "student" version does not
include (I think) the full compiler (just compiles in-core,
not producing files), and I think does not support
new foreign function linkages. I think that both
these restrictions would have to be lifted to be
viable.
My guess is that it will be something like
the free lisp + enough extra resources to include all
of maxima code + enough free space to continue to make it
interesting.
I don't know how people here feel about this.
Possible reactions
1. Yes, let's do it. It makes our lives simpler.
(We still need to understand what restrictions we
can live with.)
2. No, only open-source GPL code allowed here.
3. It's just another Lisp, along with GCL etc to support.
Do it if you want to.
4. Something else.
If you are inclined towards reaction 1, we still
need to make our list of "non-negotiable demands".
They must be such that Franz Inc's business is
placed in jeopardy. If we use this code it is
in our best interests that the company remain profitable.
RJF