[AstroPy] AstroLib vs Astlib vs APLpy vs ...

Joe Harrington jh@physics.ucf....
Fri Jun 26 20:58:54 CDT 2009


There *is* a central place to list all relevant packages: the
scipy.org list of Topical Software, in the astronomy section:

http://scipy.org/Topical_Software#head-4891fc4038123c6a923398225a728ec92ddb2780

If you just go to http://scipy.org/Topical_Software and click on
Astronomy, you get there, too.

Don't forget Russell Owen's excellent interface to DS9, which does not
do the interpolation of numdisplay nor require manual specification of
separate fifo ports on multiuser systems.

I would favor some kind of universal collection of utility software,
but there were problems with the social and release organizations of
IDLastro and SciPy that we should try to avoid.  For example, how
should we fund it, how should we review contributions, should we
choose just one approach to each problem or include multiple
solutions, should we include large packages, should we have a sandbox,
how will we coordinate releases, how do we enforce good documentation,
etc.  Already, we have several candidate collections, none of which,
to my knowledge, has gotten the community behind it nor answered any
of the "thinking big" questions.  I think we ought to do that before
moving forward with a collection.

But, I also think everyone must play in a "sandbox"-type testing
ground (or release independently) for at least a year before getting
accepted into the collection package.  To get out (and into the
collection) you must demonstrate: broad adoption for your package,
stability, maturity (i.e., a relatively slow release cycle compatible
with the big collection), reliability (low numbers of bug reports),
top-quality reference and user docs, and a commitment to long-term
maintenance.

So, people should aggressively develop software and post notices of
their existence both here on this list and in Topical Software.  The
key to a well-accepted collection will be a lot of well-tested and
well-discussed software.

Meanwhile, let's talk about a sort of charter for this collection
online, and wrap up with some decision-making at an astronomy BoF at
SciPy 2009 in August.  We can skype in anyone who can't make it.

--jh--


More information about the AstroPy mailing list