[SciPy-Dev] spectral methods in Python

[Ralf Gommers]

On Thu, Feb 14, 2013 at 3:05 PM, Nikola Mirkov <largeddysimulation at gmail.com
> wrote:

>
> Hello guys! I'm new to this list.
>

Hi, welcome.

>
> I'm very interested in spectral methods, and for that purpose I use
> Python, numpy and scipy a lot.
>
> Recently I have made an extension module for Python, consisting of some 85
> functions, useful for implementation of spectral methods.
>
> It is based on D.Funaro's Fortran library *splib*. Therefore I call my
> project *pysplib*.
>
> Here's the link to my github repo:
> https://github.com/nikola-m/pysplib
>
> The naming scheme of functions is very simple and economic (concieved by
> D.Funaro). It uses three pairs of letters denoting 1) what item we want to
> return (nodes, derivative matrices, expansion coefficients,function
> values,etc.), 2) which orthogonal functions we operate on
> (Chebyshev,Legendre,Jacobi,Hermite), and 3) what set of points we use
> (Gauss,Gaus-Lobatto,Gauss-radau). So we have:
>
> zechgl - ZEros (nodes) of CHebyshev polynomials at Gauss-Lobatto points
> or
> colega - COefficients, LEgendre, GAuss
> etc.
>
> There's a very comprehensive documentation in pdf, covering theoretical
> aspect of these functions.
>

Link: http://cdm.unimo.it/home/matematica/funaro.daniele/rout.htm


> I would like to get some feedback,so send an email if you have any
> comments or questions.
> If it proves useful for scipy as a community, you may adopt it, and maybe
> include in some module.
>

First observation: licensing seems problematic. The original code doesn't
contain a license as far as I can see. Did you ask Prof. Funaro if he's OK
with you redistributing his code under the GPL? Even if you did, GPL code
is not OK for inclusion in SciPy, it needs to be BSD-compatible.

The functionality sounds interesting though.

Cheers,
Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scipy-dev/attachments/20130216/44cb9651/attachment.html>