I have seen some stir recently about implementing a CPAN work-alike for Python. I think that Python doesn’t really need a CPAN. First off, take a look at the Python Standard Library. There have been several instances where I was looking for some python code to do something, say parse a url string into several parts. Guess what? Python has a built-in module called urlparse that does it for you.
Sure, there are things that the Python Standard Library can’t do. For that, distutils makes it really easy to distribute and install Python packages or modules. All the end user should have to do is python setup.py install. Most of the time, required libraries or modules will be included or linked to on the web or in the docs. It’s true that with distutils you don’t have a way to travel down the dependency tree to install things in the order neccesary, but with all the goodies in the Standard Library and distutils, that’s not usually an issue.
I have a feeling that a CPAN workalike would duplicate a ton of work that has already been done in the Python community. It probably also won’t help all that much.
Also, many of the modules that you might use to argue the neccesity of a CPAN-like archive should alternatively be folded in to the Python Standard Library. They’d probably be a lot more useful there.
I have a feeling that there might be some discussion about this at PyCon DC, as I think someone mentioned it on the PyCon Wiki at some point. Just a quick thought/pseudorant.