[Catalog-sig] What is the point of pythonpackages.com?

[Martijn Faassen]

Hi there,

On Tue, Feb 7, 2012 at 6:05 PM, PJ Eby <pje at telecommunity.com> wrote:
> Note, too, that prior to setuptools' development, there wasn't even any
> expectation that projects listed on PyPI even have a current *release*,
> or even have any *source code written* ,

That a package that only exists in a version control system can be
listed on PyPI makes sense to me. That a package without any source
code is listed on PyPI makes less sense to me; are there examples
where that does or did make sense?

But if at some point there *was* a release of the package listed in
PyPI and now, without author intent involved (to leave out the moral
arguments) the release cannot be found anymore, I'd say PyPI is
incorrect.

There are two possible cases:

* release cannot be found at all anymore by humans clicking around
PyPI; it's not in PyPI nor in any linked sites.

* release can be found by a determined human in linked sites, but
automated tools now fail to download the case where they did not
previously.

The first case you could argue is a flaw in the tools and not an
incorrect index, though you could also argue that tools shouldn't be
based on unreliable magic behavior, so you could argue the index isn't
providing one of the services that people rely on.

The second case is an actual incorrectness in the index.

> Just because the common use-case for PyPI nowadays is to pull down
> installation files, doesn't mean the previous use cases which PyPI catered
> to are gone or not worth supporting any more.

I wasn't arguing otherwise myself.

But I am interested in approaches that make releases that were listed
on PyPI and were accessible by automated tools permanently accessible
by such tools (unless author intent is involved; I want to leave that
off the table in this discussion).

Regards,

Martijn