Re: Why Maxima is in Lisp/ if you want to rewrite, here are some projects

>>>>> "CY" == C Y <smustudent1@yahoo.com> writes:

    CY> --- Richard Fateman @cs.berkeley.edu> wrote: > Writing an alternative front end for Maxima is a >

This is so mangled I can't read it!

    CY> talking to it? > PS. You could also look at other packages:
    CY> Matlab > comes to > mind (or its free clone), could be added
    CY> to Maxima. I think that's Octave? I've often wondered if a
    CY> merger of that project and Maxima would be a good idea - they
    CY> seem to have done well with the numerical end of things. Also,

There is matlisp (on sourceforge) that hooks LAPACK into lisp with
some matrix classes through in.  Only works with CMUCL and ACL
currently.  It should work with any Lisp that an FFI and stores
double-float arrays in exactly the same way Fortran would store them.

I've always thought that would be cool to put the two together.
Perhaps this is what the SENAC (?) guys in New Zealand did (with
different packages)?

    CY> does anyone know anything about the plplot library? In my

plplot does not work on Solaris.  Doing a foreign function interface
is a pain.

Otherwise, it looks ok.  Don't know how easy it is to use.

    CY> experience, at least, the current plotting routine seems a bit
    CY> slow under Linux (it does OK on Windows) and plplot seems to
    CY> handle a lot of the same types of operations. I guess I have

Is it because maxima is slow in generating the data for plotting or is
it because the display of the plot is slow?  The Klein bottle demo is
pretty slow on my 300 MHz Ultra 30.  Rotating the plot with the mouse
is pretty fast, but the surface is very randomly patched until I stop
moving the mouse.  Then the display is fixed very quickly.  On my 800
MHz Linux box, I think the surface rotates nicely in one piece.

I think maxima has the right idea:  Use some external process to do
the actual display and let maxima generate the data.  Doing anything
else makes maxima much less portable.

Ray