[python-committers] [Python-Dev] PEP 581 (Using GitHub issues for CPython) is accepted

[Ezio Melotti]

Hello,

On Wed, May 15, 2019 at 5:18 PM Paul Moore <p.f.moore at gmail.com> wrote:
>
> On Wed, 15 May 2019 at 15:56, Victor Stinner <vstinner at redhat.com> wrote:
> >
> > Hi Paul,
> > Le mer. 15 mai 2019 à 11:40, Paul Moore <p.f.moore at gmail.com> a écrit :
> > > Also, is there an archive of the discussions anywhere? The PEP says
> > > discussions happened on Zulip, but I don't follow that and I don't
> > > know where I can find an archived copy of the discussions.
> >
> > Well, the PEP has been discussed a lot at many places since May 2018.
>
> Thanks for all of these. I appreciate the time you took locating them
> for me. But I do have to say that I still can't really follow any
> useful "thread of discussion" - it all seems very fragmented, and it's
> difficult to see the progress towards consensus. Maybe that's just
> because I'm too used to mailing lists :-)
>

I share the same concerns:
1) the discussion was fragmented between
zulip/discuss/github/python-dev/python-committers/sprints/pycons and
very difficult to follow, even for interested people (Victor already
posted several links but missed a few others);
2) the progress toward consensus was not clear and the approval came
somewhat unexpectedly (it was mentioned a couple of weeks ago on
https://mail.python.org/pipermail/python-committers/2019-April/006705.html
and AFAICT no further discussion took place in public forums since
then);

In addition:
1) the PEP contains several factual errors.  I pointed this out during
the core-sprints last year and more recently Berker pointed out some
on GitHub: https://github.com/python/peps/pull/1013 ;
2) the "discussions-to" header of the PEP points to the zulip stream.
The stream has not been active for 6 months (it got a few new messages
today, the previous activity was in Dec 2018);
3) most of the discussions linked by Victor happened last year.
Unless I missed some, the only discussions happened this year are the
two on Discuss from February (with 3 messages each from a total of 5
authors), and the python-dev thread from March (with 12 messages from
7 authors).  One of the two Discuss threads was a inquiry about the
process (https://discuss.python.org/t/move-pep-581-discussion/873);
4) Berker is/was writing a competing PEP, in order to address the
problems of PEP 581 more effectively since his comments on GitHub
haven't been addressed;
5) next week a student is supposed to start working for the PSF on
b.p.o and Roundup as part of Google Summer of Code
(http://python-gsoc.org/psf_ideas.html);
6) PEP 8016 says "The council should look for ways to use these powers
as little as possible. Instead of voting, it's better to seek
consensus. Instead of ruling on individual PEPs, it's better to define
a standard process for PEP decision making.";

To summarize, I feel the approval of this PEP is premature and that
consensus was reached in a way that wasn't very transparent, without
considering some of the concerns.
(This might also be a symptom of a wider problem caused by the
fragmentation of the discussions between the old MLs, discuss, zulip,
IRC, GitHub PRs and issues, and IRL meetings, but this is a separate
topic.)

Best Regards,
Ezio Melotti




> > The PEP 581 has been (first?) discussed at the Language Summit which
> > was part of Pycon US 2018 (May 2018).
>
> Was that written up, or is it all just from people's memories by now?
>
> > https://github.com/python/peps/pull/681/
>
> Ah - I don't really follow this sort of PR discussion, as the github
> emails don't tend to have sufficient context on what's being said, so
> I (mostly) gave up a long time ago. Also, I tend to assume that
> discussions on PRs are mostly about details of wording, and
> substantive changes will be dealt with in a wider forum. I wonder if I
> should be following PRs on the PEPs repository more closely...?
>
> > Multiple threads on Discourse:
> >
> > https://discuss.python.org/t/move-pep-581-discussion/873
> > https://discuss.python.org/t/pep-581-using-github-issues/535
> > https://discuss.python.org/t/what-are-next-steps-for-pep-581/864
> > https://discuss.python.org/t/pep-process-after-pep-8016/558/4
> >
> > Thread on python-dev:
> >
> > https://mail.python.org/pipermail/python-dev/2019-March/156626.html
> >
> > Threads on python-committers:
> >
> > https://mail.python.org/pipermail/python-committers/2018-May/005428.html
> > https://mail.python.org/pipermail/python-committers/2018-June/005506.html
> > https://mail.python.org/pipermail/python-committers/2018-July/005657.html
>
> I saw these, but didn't get much of a sense of progress towards
> agreement. Again, maybe just because they were lots of fragmented
> threads and locations.
>
> > Discussion on Zulip Chat:
> >
> > https://python.zulipchat.com/#narrow/stream/130206-pep581
>
> I can't see this without logging in :-(
>
> > The Steering Council discussed it internally as well:
> >
> > https://github.com/python/steering-council/blob/master/updates/2019-04-26_steering-council-update.md
>
> I did see that, I was more wondering what led to that decision
> (whether it was a general consensus from the core devs that it was a
> good move, or mainly the SC's own view that prevailed).
>
> > The PEP 581 and 588 have been discussed at the Language Summit which
> > was part of Pycon US 2019 (2 weeks ago).
>
> Again, has there been any write up of that (yet)?
>
> As I say, I don't object to the decision, I'm more just trying to
> better understand the process of being involved under the new regime
> of the SC, combined with multiple fragmented forums for discussion. It
> feels a lot harder these days to keep track of all the
> discussions/decisions going on. But maybe that's a good thing - only
> people with a genuine interest get involved, and I can spend less of
> my time reading mailing lists! :-)
>
> Paul
> _______________________________________________
> python-committers mailing list
> python-committers at python.org
> https://mail.python.org/mailman/listinfo/python-committers
> Code of Conduct: https://www.python.org/psf/codeofconduct/