[AstroPy] Proliferating py-astro-libs
thomas.robitaille at gmail.com
Thu Jun 9 22:58:39 EDT 2011
> I agree mostly with what Marshall says, but as a (sort of) an
> outsider, it's still quite obscure what projects are considered to be
> the "main" community efforts. I still don't really know which
> packages should be the de facto standards to do certain things for my
> own work (aside from PyFITS, PyRAF, stsci_python... something that
> comes out of STSci).
> My point is, if you claim there *are* community efforts that we could
> contribute to, we would appreciate which ones they are and advertise
> as such.
I don't think it matters too much if a package is *the* main one, but as a rule of thumb, it'd be better to contribute to *any* of the existing packages than to write a new one. So if there are three WCS libraries, then it's best to contribute to the one you prefer than to say that none are the 'main' one and create a new one!
Another rule of thumb is that if a package is on GitHub, then it's very easy to contribute to it via pull requests, and developers really appreciate it when users fix bugs or add features this way. Packages hosted on custom-hosted SVN servers are more difficult because the only mechanism in place is to send a patch by email and the process is much less transparent and your email might just get ignored.
Maybe we need some kind of voting system where users can list what packages they use, and it would then be more obvious which are the most widely used ones? (like http://osx.iusethis.com/). Any volunteers to set up such a system?
More information about the AstroPy