F24 Change for PDC

[Josh Boyer]

On Fri, Nov 6, 2015 at 1:37 PM, Matthew Miller <mattdm at fedoraproject.org> wrote:
> On Tue, Nov 03, 2015 at 11:46:52AM -0500, Josh Boyer wrote:
>> > https://fedoraproject.org/wiki/Changes/PDC
>> This sounds fantastic from a "should we do this" standpoint.  However...
>> > Anyone want to give it a review, make any edits, or add yourself to it
>> > before I send it in?
>> Is there a reason you wrote it up as a Change?  Normally FESCo reviews
>> those and approves them.  Given that this is entirely self-contained
>> in Rel-Eng and Infrastructure, I'm not really sure a) there's anything
>> FESCo needs to review here and b) it's even FESCo responsibility to
>> approve it.
>
> So, for whatever it's worth, I find it really helpful to have these
> documents for infrastructure improvements, to help advertise the
> planned change and to make sure any stakeholders have a chance to
> provide early input. And often the process of putting together a
> planning document as simple as the one required for Changes brings
> clarity to the actual tasks that need to be done, so it can be helpful
> for its own sake.

Yes, sure.  I don't disagree with any of that.

> Since it's a self-contained change, as you say, it just goes into the
> big bucket of those and normally gets a rubber stamp, right? Unless
> FESCo decides that there _is_ something special to review, and then
> there's a process for that.

My concern is that FESCo is not the body to actually decide if there's
something special here.  Infrastructure FESCo members aside, that body
does not have expertise or insight into how the backend of things is
run.  So if rel-eng is hoping that FESCo is going to catch something
that needs further review, then I am afraid that is misplaced
optimism.  I don't think there's anything wrong with more eyes, but I
question whether the process here is really doing to do anything
besides a rubber stamp.  Particularly when self-contained changes are
usually approved wholesale.

(And while it is a large tangent, the fact that we knowingly have a
process that is heavily a rubber stamp indicates that process isn't
providing value as-is for such changes.)

I would hope instead that rel-eng and infrastructure would heavily
review their own changes and somehow indicate on the document that
multiple people from each group have signed off on them.  If we're
going to stick with FESCo as the "approval body", then at least having
that list of reviewers from each group shows that rel-eng and infra
are mostly in agreement.

josh