On Tue, 2011-02-15 at 06:49 +0200, Peter Eisentraut wrote:
> On mån, 2011-02-14 at 11:49 -0500, Stephen Frost wrote:
> > Perhaps a thought for next time would be to offset things a bit. eg:
> >
> > CF 2011-03 (or whatever):
> > 2011-02-14: Patches should all be submitted
> > 2011-02-14: Reviewers start
> > 2011-03-01: Committers start w/ 'Ready for Committer' patches
> > 2011-03-14: Patches not marked 'Ready for Committer' get bounced
> > 2011-03-31: All patches committed
> >
> > I'm not against the 'waiting on author' approach, but I do feel like
> > if we're going to continue to have it, we need to spread it out a bit
> > more.
>
> I don't think it is realistic to add even more dates and bounds and
> guidelines.
Though I agree with that, I see Stephen's suggested sequence of events
as a useful one for CFs managers.
> People are already widely ignoring the current ones.
I don't think people are ignoring things deliberately. List traffic is
high and unless you are 100% full time on this, you can't possibly hope
to read them all, respond to the relevant ones and do your own work too.
Especially when many folk have a day job as well.
> If you want to have the ability the bounce things more aggressively, I'd
> argue for shorter and more frequent commitfests. Say, one week per
> month.
That is the blink of an eye for me, so don't want more frequent CFs.
The focus should be on organising ourselves so that the most number of
high quality features get into Postgres. I don't see anything to be
gained by bouncing things aggressively; strict time-boxing works on 2
weekly cycles, not on annual ones.
We're doing OK.
-- Simon Riggs http://www.2ndQuadrant.com/books/PostgreSQL Development, 24x7 Support, Training and Services
