C Y wrote:
>
> --- David Holmgren <DavidHolmgren@smarttech.com> wrote:
> > Hi - I intended to write logexpand:super , which does solve the
> > problem.
> > Generally speaking, with most computer algebra systems you have to
> > give the
> > program some assistance at some stage in the calculation. I think
> > that this
> > is the point that Dr. Fateman was trying to get across; you have to
> > be
> > willing to use things like expand and factor.
>
> Well, I don't really have a clue about stuff like this, but why
> couldn't we teach the solve package to try some of the expand and
> factor stuff itself? If it doesn't like the form that's given, try
> changing it around a bit before abandoning the effort?
Since you seem to think this is easy, feel free to write such
a program! Note that some of the programs that you might try
(a) destroy some of the solutions
(b) create extraneous solutions
(c) can take extremely long times
(d) explode a problem into huge size
(e) still might not help.
Note that almost any computation done by a computer algebra
system could be thought of as
solve(<something>) and therefore you have almost no
guidance as to what is really required as an answer.
Numerical approximation?
Solution over the complex numbers?
Taylor series?
Algebra or differential or difference or .... equations?
Do you want solve to be able to do this:
solve(f(x)^2-f(x)=0,f) ?
PS. It is unlikely that solve likes or does not like forms
of input :)
You could try writing what amounts to a bunch of rules,
as a first stab at it.
RJF
>
> __________________________________________________
> Do You Yahoo!?
> Listen to your Yahoo! Mail messages from any phone.
> http://phone.yahoo.com
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima