Re: Cannot use ldegree() on symbolic expressions with non-integer exponents

[Martin Helm]

Am Samstag, 27. März 2010 18:04:05 schrieb Martin Helm:
> Am Samstag, 27. März 2010 17:23:23 schrieb Jason Lewis:
> > Thanks for the reply.
> >
> > It appears that ^ doesn't like 'int32' as an argument, and so it
> > produces an error.
> >
> > I've run into this problem while trying to have octave solve a system
> > of symbolic equations to model the output of an electronic circuit
> > (something I used to do with Matlab). I'm a bit disappointed, but I'm
> > willing to concede that I may be the problem.
> >
> >  - J
> 
> What about using a CAS for this. The reason I do not use the symbolic
>  package is simply that I use a full CAS (in my case maxima) whenever I
>  need to do symbolic calculations. Numerical computations are of course
>  much slower in maxima, but it depends on the rest of what you do if this
>  will be a problem or not.
> 
> Nevertheless you should post your question on the octave forge mailing list
> (it is a problem with a package from octave forge not an octave problem).
> Maybe the solution is absolutely trivial and I simply do not know it.
> 
> Seems there are not many people hanging around here using this package.
> 
> Btw one last shoot in the dark, does it accept something like
> y = sym("x^2 + 9") ?
> 
> - mh
> 
I installed now symbolic package  1.0.9 with octave 3.2.4 and played a little 
bit with it and the examples I found on the net (cheb poly and newton method) 
and tried your problem and compared it to the matlab examples for ldegree.
Really nice.

I would say this is simply a bug in ldegree and degree, report this to octave 
forge (not to the bugs at octave mailing list).

If you have luck one of the forge maintainers can and will fix it.

- mh