Not enough Python library development [was PEP scepticism]

[Bruce Sass]

On 4 Jul 2001, Ralf Muschall wrote:
> Grant Griffin <not.this at seebelow.org> writes:
> > concept seemed to encourage large interdependencies between modules.
>
> Unfortunately, the logical alternative to interdependencies is
> duplication of code (or, at least, of functionality).
>
> > "application-targeted" distributions.  For example, someone could make a
> > "Numeric" bundle of Python, including NumPy and related modules;
> > likewise, there could be a "webmaster" bundle.  Of course, this takes a
>
> This does not remove the need to keep them mutually compatible.

Stuff would _need_ to be packaged in terms of the smallest replaceable
unit if you want to eliminate duplication, and if you do not eliminate
duplication you can end up with packages conflicting because they
require different versions of the same software.

One result of this is that packages which make it into the archive,
and depend on other stuff in the archive, will either get maintained
or dropped when the dependencies can not be met anymore.

The bug tracking system is what tells the pkg's maintainer there is a
problem, but if the maintainers are not around because you didn't make
them jump through some hoops to ensure they are serious -- it all
falls apart and becomes easier for the user to just go get the source.


I think this needs a PEP from the core Python developers,
for four reasons:

1) anything less is "just" third-party (no matter how well done)

2) to end up with an archive that is not crufty, inconsistent, or
   full of duplication -- Python itself must be in the archive
   [(co)authoring a PEP is tacit approval]

and

3) having a repackaged version of Python in the archive is
   potentially starting down the road to having multiple,
   incompatible, archives (who decides the right way to break
   Python down into packages, and what happens when PyPAN and
   ActiveState (or whoever) diagree) -- this pretty much
   necessitates the core developers involvement

4) this is a complex undertaking, the result of which can affect
   the stability and security of systems that use it


- Bruce