Re: [gentoo-project] Call for agenda items - Council meeting 2014-04-08

[Rich Freeman]

On Sun, Mar 30, 2014 at 7:00 AM, Anthony G. Basile <blueness@g.o> wrote:
> While implementing a new policy may not be the right approach (or so I'm
> hearing from the community), I can bring forward at least 3 examples of
> significant changes that were not discussed.  I don't think I would have
> difficulty convincing people of this fact.  If we do not enact policy then
> how is this problem addressed?

I should clarify my meaning.

I don't think we need to enact a specific virtual/profile/etc policy
because I think it should ALREADY be treated as policy.  At most we
should be clarifying that we already consider it policy.  To the
extent that we create new policy it should be much more open-ended,
like "consult the list when making major design changes that impact
many packages/users" or something like that.

However, either creating/clarifying policy, or pointing out that
something already is policy will not do anything if people don't
follow it.

I also was not suggesting that we should not enforce policy - only
that doing so is tragic.  I do think that allowing productive
developers to just ignore the rules is more harmful than stepping in.

I will make the general statement that if people have a problem with
QA outright defiance is something that should almost certainly lead to
a ban of some kind.  There are many ways to handle perceived abuse of
QA power, and that is probably the worst possible.  No need to just
sit on your hands until the next council meeting - complain privately
to council, or publicly to council, or go to comrel (who will probably
just end up arbitrating or handing off to council) but don't
unilaterally get into a revert war with QA!  When you do, you
immediately change priority #1 from fixing the situation to fixing
you.  Don't just assume that QA won't back down or the Council won't
step in after just a bit of reasoning.

On the other hand, these are people issues, and when a need for
enforcement comes up somebody should at least take the time to chat
with the individual concerned and try to explain/connect/etc, and get
the full story before taking action.  They shouldn't just get an email
saying "the Council just met and FYI your commit access is revoked for
two weeks, have a nice day."  Somebody should be responsible to reach
out to them and be a contact through the experience.  Some may
ragequit and be lost all the same, but really the goal is to try to
teach a lesson and some kind of mentoring will help with that, and
perhaps give the affected individual someplace to vent the next time a
problem comes up.

Rich