Xref: utzoo comp.sources.wanted:16815 sci.math.symbolic:2496
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!samsung!mitech!gjc
From: gjc@mitech.com
Newsgroups: comp.sources.wanted,sci.math.symbolic
Subject: Re: Software for handling *simple* polynomials?
Message-ID: <66@mitech.com>
Date: 23 May 91 14:59:03 GMT
References: <1991May22.010552.10306@jarvis.csri.toronto.edu> <1991May23.142627.5912@fys.ruu.nl>
Organization: Mitech Corporation, Concord MA
Lines: 26

In article <1991May23.142627.5912@fys.ruu.nl>, spit@fys.ruu.nl (Werenfried Spit) writes:
>> In sci.math.symbolic you write:
>> 
>> I'm currently playing around with some polynomials that appear in a
>> doubly-recursive relation (I may post details about it if I get too
>> frustrated).  I have been using Maple, but it is running out of swap
>> space.
>> 
>  How about good old REDUCE? 
>  Seriously; I have spent some time in integrating Gaussian like
>  functions symbolically; this turns out to be manipulation of 
>  polynomials with arbitrary coefficients. I tried it with 
>  Reduce, with Maple and with Form...

One reason Maple might be having problems here is that Maple uses a technique
that trades off SPACE for TIME by keeping a table of previous computation
results. (Maybe there are ways of turning this off?)

Another "good old" system that might work here is Macsyma, which
has a representation called CRE that makes some polynomial
operations quite a bit faster. 

Also, with Macsyma it is possible to take a user-written procedure and
compile it into MACHINE LANGUAGE. *Sometimes* this helps a lot.

-gjc