[AstroPy] Co-ordinating Python astronomy libraries?

[James Turner]

I'm still catching up on reading this monster thread between doing
vacation things (thanks for all the comments, everyone)... I'll
quickly comment on this first.

> The Fortran version of SLALIB that I wrapped was GPLd.  You can get the 
> code here (SLALIB + wrappers and tests):
> 
> http://www.cv.nrao.edu/~sransom/pyslalib-1.0.tar.gz

Yes, the latest SLALIB does have an open source licence, but it's GPL,
which will put off a lot of the SciPy community (which may or may not
matter in this case). Any application that links with SLALIB will also
need to be GPL (at least its binaries) and the SciPy people are quite
adamant about using BSD. I suspect that's largely because Enthought
has contributed a lot to SciPy and their business model depends on
selling the code in commercial apps. It's hard to argue with that,
given how central the Enthought people have been in making SciPy and
NumPy happen.

There is, however, an older version of SLALIB distributed with IRAF,
prior to JAC relicensing the library with their changes under the GPL.
The licensing of that is not clearly stated though, as far as non-IRAF
apps are concerned.

Does anyone have comments on 1) whether people need the latest SLALIB
(or is IRAF's version good enough?) and 2) whether we're concerned
about licence-compatibility with SciPy for libraries that are so
astronomy specific? For more general stuff (eg. deconvolution) I'd
suggest that we should care about the licence, so we can share code
with other disciplines and maintain goodwill in the SciPy community.

Cheers,

James.