Re: Multivariate polynomails

[Ben Abbott]

On Apr 13, 2013, at 4:25 PM, Juan Pablo Carbajal wrote:

> On Sat, Apr 13, 2013 at 10:19 PM, Ben Abbott <address@hidden> wrote:
>> On Apr 13, 2013, at 1:25 PM, Juan Pablo Carbajal wrote:
>> 
>>> Hi all,
>>> 
>>> What do you think of this package?
>>> 
>>> It is not fully shaped as an Octave package but maybe we can add it to 
>>> Agora?
>>> 
>>> Is there a multivariate polynomial package for GNU Octave already?
>>> 
>>> Thanks
>> 
>> Juan, which package?  Missing a link?
>> 
>> Ben
> 
> Indeed, my apologies.
> Here is the link
> http://soliton.vm.bytemark.co.uk/pub/jjg/en/code/mvpoly.html
> 
> I do not know if there is a better package for multivariate
> polynomials. Maybe we can convince this person to continue developing
> the package.
> At the moment is somehow useful but the conventions used are different
> from the standard polynomial coefficient storage.
> 
> Thanks for the heads up.


Regarding the polynomial coefficient convention, I've worked with 2D 
polynomials in the past and found that a particular convention was useful when 
overriding polyder() and polyint().   I haven't looked at extending polyval, 
polyder and polyint to more than two variables, but my impression is that a 
sparse ND array representation would be a good solution.

Ben