I like the idea a lot, but the main obstacle lies in migrating a
"critical mass" (which I can't define) of the significant code
libraries over to code4lib -- enough so that the repository becomes
recognized as valuable resource. Following on that, would the site be
a mirror site (and what sort of permissions from code owners does this
entail) or are we expecting the various project leaders to be
responsible for maintaining up-to-date versions on this site and
perhaps their own.
I also wonder about the relationship of the code4lib repository with oss4lib.
Obviously, oss4lib is a different sort of site -- making links to
software projects rather than hosting the code itself. However,
harking back to my first point, my experience with oss4lib is that it
misses important projects and that diminishes its usefulness for me.
For example, to use Gabriel's example, pymarc is not included.
I'd be interested in hearing from Dan about his reaction to this idea
and especially about the relationship of oss4lib and code4lib.
On 8/13/07, Gabriel Farrell <[log in to unmask]> wrote:
> In #code4lib today we discussed for a bit the possibility of setting up
> something on code4lib.org for code hosting. The project that spurred
> the discussion is Ed Summer's pymarc. The following is what I would
> like to see:
> * projects live at code.code4lib.org, so pymarc, for example, would be
> at code.code4lib.org/pymarc
> * svn for version control
> * trac interface for each
> * hosted at OSU with the rest of code4lib.org, for now