[Python-Dev] Consolidating channel of record [was: 581 (Using GitHub issues for CPython) is accepted]

Paul Moore p.f.moore at gmail.com
Wed May 15 18:07:19 EDT 2019

On Wed, 15 May 2019 at 21:34, Chris Barker - NOAA Federal via
Python-Dev <python-dev at python.org> wrote:
> Frankly, multiple long meandering threads in s single mailing list are
> not s very good archive either.

They aren't, I agree. But in my view, they constitute the *source
material* that the summaries in the PEP should be based on. No-one
should need to read the archives if they just want to know the points
being made. But for someone who wonders how a particular conclusion
was reached, having a single archive of record (to use Stephen's
phrase) makes sense - certainly people *can* search multiple sources,
but we should be trying to be transparent, and not making it too hard
to research the background of a decision is part of that.

> Ideally, the PEP is updated with a summary of the issues discussed as
> the discussion unfolds.

I'm not sure that's just "ideally" - one of the key purposes of a PEP
should be to summarise the discussions.

> (And links to the main discussion threads?)

That's nice to have, yes.

> It founds like that didn’t happen in this case, but it’s not
> necessarily too late.

Definitely. In fact, I doubt there's much of any real interest that
needs to be referenced (workflow debates are not the most interesting
reading, in my experience :-))

> As the community seems to be moving to a wider variety of fora, this
> will become all the more critical.

I completely agree here.

> On May 15, 2019, at 1:09 PM, Stephen J. Turnbull <turnbull.stephen.fw at u.tsukuba.ac.jp> wrote:

> > Please, let's not start by privileging any particular type of channel
> > in this discussion.  I know what I like, but it's far more important
> > to have a single place to refer to past discussion IMO.  It's bad
> > enough with python-ideas and python-dev.

100% agree. This shouldn't be about whether any particular channel is
"better" or "worse" than any other. It's just about not having the
original content of the discussions that feed into an accepted PEP be
scattered too widely.

> > The stricture for the Council deliberation channel is different, since
> > I expect the archives would be private to Council members: if you came
> > into this discussion in the middle, what conversations would you want
> > to be able to review?

In theory, PEP 13 says that the council should look for consensus
rather than making decisions on their own. Ideally, the council's
private deliberations should be of little interest to anyone else,
because either (a) the consensus should be clearly visible from the
public record of the debate, or (b) the public record should show that
the discussion was getting nowhere, and the council had to make the
final decision (at which point, the community has effectively given up
the right to argue about why the council chose what they did).

The reason the decision on PEP 581 started me thinking about this was
precisely because I hadn't seen any signs of that consensus or of a
stalled discussion. Maybe I wasn't looking in the right places, but I
*thought* I was following the majority of PEP-related discussions (at
least skimming them). So I was somewhat startled to see a formal
decision come with what felt like no real warning.


More information about the Python-Dev mailing list