On 6/22/07, Reinhard Oldenburg <oldenburg at ph-heidelberg.de> wrote:
> Is there a Lisp implementation for the Java VM that can compile Maxima?
> It seems that armed bear common Lisp may be a candidate. Has anyone
> manged to compile maxima on this Lisp?
I have managed to compile Maxima with Armed Bear CL.
The resulting program is indeed intolerably slow
(much slower than Clisp, which the slowest widely-used
CL implementation that I know of) and it failed Maxima's
test suite. The ABCL project appears to be dead, so,
unless someone takes over from the original author, I doubt
if the remaining bugs will ever be fixed.
I think the most promising avenue to run Maxima on a JVM
is to build a Java bytecode generator for Clisp.
That is an item on the Clisp wish list but there has been no
progress on it. I'm pretty sure building a bytecode generator
for an existing CL implementation is about 1/100 times the
effort of rolling a new implementation, so I'll take this
opportunity to encourage people looking for something to do
to take up the Java bytecode generator for Clisp.
The idea of Java + Maxima comes up from time to time.
You can search the mailing list archive
(http://maxima.sourceforge.net/maximalist.html)
to see where the conversation has gone before.
HTH
Robert Dodier