On Tue, Oct 4, 2016 at 5:15 PM, Graham Knop <ha...@haarg.org> wrote:
> Forgive me if I don't.
>
That's fine. Your comments address the meta-issue.
> PAUSE has no governance model.
Nor, really, does "Perl" in the broadest sense. PAUSE admins have
discussed this in coming up with a response. What people think of as
"Perl" or the "Perl ecosystem" is a loose federation of systems that work
because the owners/maintainers of those systems discuss and work together
and generally trust each other to act in the interests of the whole. In
this model, power derives from administrative control, checked by the
powers of other administrators controlling related parts. (E.g. Perl NOC
controls certain servers, but Elaine owns the cpan.org domain.)
There is probably a good blog post to write about this, which either
someone else will write or I'll get to once I'm not busy with current
affairs.
> The actions taken by the PAUSE admins in this recent dispute, assuming
> they are fully represented by David Golden's emails,
Let me assure you that everything I've done has been in consultation with
other administrators and I continue to update them and discuss the
situation as it evolves.
While the initial message states ["PAUSE administrators
> would like you to talk this out"][1], it is clear from further emails
> that [the PAUSE administrators are in fact *demanding* ("acted by fiat")
> that this be talked out.][2] They are predicating their decision on the
> community. So it has been stated that [the community will override
> PAUSE's permissions model][3], and essentially, RIBASUSHI's first-come
> permissions have already been revoked.
>
>
The difference between #1 and #2 is my fault. I originally intended a more
direct statement and let myself be swayed to a more polite request. My
original intent wouldn't have had that gap.
Peter's claim on unrestricted first come permissions are contradicted by
his original "verbal" (IRC) agreement with Matt about the nature, purpose
and duration of the transfer. Peter's argument is that his subsequent work
gives him moral authority to disregard that agreement.
PAUSE administrators feel the best people to assess the moral authority
argument are the people already involved: the maintainers and ultimately
the community.
Particularly, as there had been no discussion of any specifics of Peter's
plan, arbitrating in favor of any side seemed incredibly premature.
Therefore, as stated, we viewed the best course of action to be to slow
down the process (the "no unilateral transfer" part) and encourage
communication to see if the problem can resolve itself without further
administrative fiat.
This we have done and continue to think is the course of action that is
best for the community.
On what grounds is the first-come permissions holder and primary
> maintainer being circumvented? Does this apply to other modules?
I've explained the grounds above. I don't know of any other module with a
similar "river position" with a similar transfer agreement. We don't
intend for this dispute to set a broader precedent for disputing namespace
permissions generally.
Personally, I agree that a third permissions level that distinguishes
"ownership" from "co-maint maker/unmaker" would be helpful and hope that's
something that can be put in place in the future.
David
--
David Golden <x...@xdg.me> Twitter/IRC/GitHub: @xdg
