Newsgroups: sci.math.symbolic
Subject: mma.src
Summary: Programs that simulate Mathematica are available.
Expires: 
References:
Sender: Richard Fateman
Reply-To: fateman@peoplesparc.Berkeley.EDU (Richard Fateman)
Followup-To: 
Distribution: 
Organization: UCB CS
Keywords: mathematica Common Lisp

Faithful readers of this list may be aware of this program, but
it's gotten somewhat more capable, and can now be copied for
your own use.  Here's an example of what it can do:
......
<cl> (tl) ;; top level program


In[1] := Timing[RatExpand[(x+y)^4]]

                       4      3        2  2        3    4
Out[1] = {0.0 Second, x  + 4 x  y + 6 x  y  + 4 x y  + y }
In[2] := Timing[RatExpand[(x+y+z)^15];done]

Out[2] = {0.23300001 Second, done}
In[3] := D[ArcTanh[x]^2,x]

         2 ArcTanh[x]
Out[3] = ------------
                 2
            1 - x
In[4] := Int[%,x]

                     2
Out[4] = (ArcTanh[x])
In[5] := Int[x^n,x]

                    n
Out[5] = integrate[x , x]
In[6] := integrate[y_^m_,y_]:= (y^(m+1)-1)/(m+1)

In[7] := Int[x^n,x]

               1 + n
         -1 + x
Out[7] = -----------
            1 + n
In[8] := Exit

Exited to Lisp

t 
<cl> 
....
That line [2] seems to be rather faster than Mathematica on the same
machine. The lines 3-7 illustrate the primitive "derivative-divides"
integration package, and the possibility of enhancing it by
pattern matching.
....

To reconstruct this on your unix system (I'm using Allegro Common Lisp,
and it should work on any hardware supporting it; it should also work
in other Common Lisps), do the following:

cd someplace with room for a few 100k bytes

Just download the directory mma1.6  from the site below:
look for mma1.6.tar.gz
You will have to uncompress (gzip) and untar them on 
unix; on windows you can use winzip.

If you want individual files, not compressed or archived,
look in mma1.6

The site:

www.cs.berkeley.edu/~fateman/mma1.6.tar.gz 
  or
www.cs.berkeley.edu/~fateman/mma1.6/ 
   for the directory of files.
(That is version 1.6)

%% you should probably try to compile them.

.......

If you get a copy, I would appreciate e-mail (to
fateman@cs.berkeley.EDU) advising me of that fact, so I can keep you
up on new features or improvements. Of course I'd like to know of any
problems or successes you have.  On the other hand, I can't promise to
fix any particular problems, and there is no warranty.  If you are
interested in enhancing the behavior of this system for use by others,
please tell me.

Important note:   This program  is a rough shell
that has only rudimentary (but fast) parsing, display, numeric evaluation,
polynomial/ rational simplification, integration, and some pattern matching. 
 It is not a substitute for Mathematica (r), Maple (r), etc.
generally, and is not intended
to compete with such commercial systems.  This system
can be used for experimentation ... e.g. insert an entirely
different evaluation strategy  or an entirely different numeric
system, and re-run your "Mathematica(r) source code" through it.
Or hook it up to your favorite other semantics for solve, integrate,
etc.  (e.g. JACAL, Reduce, Macsyma, Maple, MuPad, Axiom ...).

It is difficult or impossible to do these experiments
with the commercial Mathematica.  WRI's initial
legal concern with the availability of an alternative
"Mathematica" parser seems to have faded.

Richard J. Fateman
fateman@cs.berkeley.edu   510 642-1879