Re: Some other symbolic math programs...

["Richard B. Kreckel" <kreckel@ginac.de>]

On Tue, 29 Feb 2000, George wrote:

> On Tue, Feb 29, 2000 at 10:38:19AM +0100, Richard B. Kreckel wrote:
> > I don't know where DrGenius is heading to but let me comment on the use of
> > such packages.  I have been considering this issue for incorporating one
> > of them into the GiNaC framework and came to the conclusion that it would
> > be better not to do so.  The problem is: All those packages are designed
> 
> My idea was to actually use GiNaC here?  Though I don't know much about it so
> I dunno how appropriate it would be.  If I understand GiNaC and am not
> mistaking it with something else, we should be able to build an equation tree
> from our parsed expression, then pass it to ginac to play with, then get it
> back and convert to our type of tree.  So here I assume ginac is a library
> for symbolic math and just happens to use cln as the package for storing
> numbers.  Am I correct?

Yes, you are correct.  The fact that both GiNaC and DrGenius use CLN would
make certain conversions very easy and unsusceptible to errors.  Mapping
the trees from one package to the other will be trivial AFAICT.  Of course
the discussion of dependencies will emerge again.  I think Cebix has done
some packages for RedHat already and I am going to pack it for Debian
Woody soon.  Maybe this helps people.

> We're not interested in speed really.  It's mostly for education and testing
> things.  If someone wants to do a lot of heavyweight calculations he'll be
> better off choosing a specific package for the specific problem.  So I mostly
> want it to do lots of stuff, rather then do limitted things fast.

Of course it all depends on what you want to do.  Maybe you should have a
look at <http://www.ginac.de/ToDo.html>, it tells you what it cannot do
(yet).  This is sometimes more helpful than a list of cool features.  ;-)

Regards
     -rbk.
-- 
Richard Kreckel
<Richard.Kreckel@Uni-Mainz.DE>
<http://wwwthep.physik.uni-mainz.de/~kreckel/>