PEP 382: Namespace Packages

["Martin v. Löwis"]

> I've maybe missed some point, but doesn't the PEP requires
> coordination so that *.pkg files have different names in each portion,
> and the same if one want to provide a non empty __init__.py.

To some degree, coordination is necessary. However, the PEP recommends
that you use <distribution>.pkg as the name; IMO, that should be
sufficient (at least when all competing pacakges are on PyPI, which
requires unique distribution names).

>> Allowing each system package to contain its own .pkg or .nsp or
>> whatever files, on the other hand, allows each system package to be
>> built independently, without conflict between contents (i.e., having
>> the same file), and without requiring a special pseudo-package to
>> contain the additional file.
> 
> As said above, provided some conventions are respected...

Yes, however, these are easy to achieve. If a conflict is ever
encountered, the author of the package violating the convention
is asked to follow it, and he usually will - or a fork will occur.

> Another point: I don't
> like .pth, .pkg files. Isn't this pep an opportunity to at least unify
> them?

I don't see this as a problem. I can add it to the discussion section if
you want.

Regards,
Martin