Enforcing the 2nd maintainer rule, was: Bug#444368: ITP: dvd95 -- DVD9 to DVD5 converter

[Reinhard Tartler]

On Wed, Jun 5, 2013 at 7:28 PM, Felipe Sateler <fsateler at debian.org> wrote:
> On Wed, Jun 5, 2013 at 12:23 PM, Reinhard Tartler <siretart at gmail.com> wrote:
>> On Wed, Jun 5, 2013 at 9:55 AM, Fabian Greffrath <fabian at greffrath.com> wrote:
>>> Am Dienstag, den 04.06.2013, 07:37 +0200 schrieb Reinhard Tartler:
>>>> It still seems to lack a 2nd person to back up the package in the team.
>>>
>>> Hm, I question the purpose of this rule. It seems to me to keep useful
>>> packages out of the archive; packages that I do not sign up for as an
>>> Uploader, because I do not know about their existence, because they are
>>> not in the archive...
>>
>> Well, it depends what you want pkg-multimedia to be.
>>
>> A package without a second supporter is de facto not team maintained,
>> but maintained by a single person. You do not need a team for such
>> packages, on contrary, they just add overhead to the team (as in, PET,
>> mailing lists, team RC bug count, etc.). In these cases, I think you
>> are much better off with keeping them in collab-maint.
>>
>> My motivation for enforcing this rule is to avoid pkg-multimedia
>> becoming a specialized QA Team. I mean, seriously, if you want that,
>> then please put "Debian QA Team" as maintainer.
>>
>> Also, please consider the bus factor. Imagine some pkg-multimedia
>> member, who has introduced 20 packages to pkg-multimedia, gets hit by
>> a bus. What is the team supposed to do with the packages? Since nobody
>> else didn't even bother to put himself as uploader, it is quite likely
>> that his 20 packages end up unmaintained. Again, orphaning the package
>> seems like a good answer to that, which in this case is unlikely to
>> happen since de jure, the package is labeled as "team maintained",
>> although de facto, nobody cares for it. That's why consider packages
>> without 2nd uploader as harmful to the team.
>>
>> I would therefore suggest to stage packages without maintainer in
>> collab-maint, and as soon as a 2nd pkg-multimedia member agrees to
>> support it, just move the repo to pkg-multimedia, and good.
>
> I agree with the principle behind the rule, but perhaps the rule is
> not the best way to enforce the principle?
>
> Brainstorm follows, posibly lousy idea:
>
> Change the rule to say that packages maintained by the team cannot
> have less than 2 uploaders, but first uploads are allowed to have only
> one uploader. This could help in breaking the vicious loop presented
> by Fabian.

Hm, I think we already have packages in the team than we can handle,
and I fear that this rule bears the risk of making the matter even
worse. OTOH, we also must not stall development and new packages! So
I'm wondering, is the 2nd maintainer rule really a problem? TBH, I
think that preparing the package in collab-maint, and asking
pkg-multimedia for uploading the package is in no way worse compared
to having the package having in the team, and finding no uploader.

Please don't misunderstand me, in no way I want to hinder great
packages to enter Debian. My concerns are towards the reputation of
team pkg-multimedia.

> Then, have a daily cron job scan the archive for packages maintained
> by the team, and flag packages with less than 2 uploaders older than N
> days. Send the resulting list to the team list.
>
> If nobody wants to help comaintain a package for M days, the package
> should be moved to collab-maint or some other area.
>
> The principle is similar to the wnpp mails sent to -devel.

I find this a very interesting idea and would welcome its
implementation with or without enforcing the 2nd maintainer rule!

Cheers,

-- 
regards,
    Reinhard