[openstack-dev] [TripleO] Spec process

James Slagle james.slagle at gmail.com
Tue Jul 29 20:54:12 UTC 2014


Last week at the TripleO midcycle we discussed the spec process that
we've adopted. Overall, I think most folks are liking the specs
themselves. I heard general agreement that we're helping to tease out
issues and potential implementation disagreements earlier in the
process, and that's a good thing.  I don't think I heard any contrary
opinions to that anyway :-).

The point was raised that we have a lot of specs in review, and
relatively few approved specs. All agreed that the time to review a
spec can be consuming and requires more commitment. We proposed asking
core reviewers to commit to reviewing at least 1 spec a week. jdob
emailed the list about that:

http://lists.openstack.org/pipermail/openstack-dev/2014-July/040926.html

Please reply to that thread if you have an opinion on that point.
Everyone at the midcycle was in general agreement about that
commitment (hence probably why not a lot of people have replied), but
we wanted to be sure to poll those that couldn't attend the midcycle
as well.

The idea of opening up the spec approval process to other TripleO core
reviewers also came up. Personally, I felt this might reduce the
workload on one person a bit and increase bandwidth to get specs
approved. Since the team is new to this process, we talked about
defining what it means when a spec is ready to be approved. I
volunteered to pull together a wiki page on that topic and have done
so here:

https://wiki.openstack.org/wiki/TripleO/SpecReviews

Thoughts on modifications, additions, subtractions, etc., are all welcome.

Finally, the juno-2 milestone has passed. Many (if not all?)
integrated projects have already -2'd specs that have not been
approved, indicating they are not going to make Juno. There are many
valid reasons to do this: focus, stabilization, workload, etc.

Personally, I don't feel like TripleO agreed or had discussion on this
point as a community. I'm actually not sure right off (without digging
through archives) if the spec freeze is an OpenStack wide process or
for individual projects. And, if it is OpenStack wide, would that
apply just to projects that are part of the integrated release.

I admit some selfishness here...since I have some outstanding specs.
But, I think we need to come to a consensus if we are going to have a
spec freeze for TripleO around the time of other projects or not, and
at the very least, define what those dates are. Additionally, we
haven't defined or talked about if we'd have an exception process to
the freeze if someone wanted to propose an exception.

-- 
-- James Slagle
--



More information about the OpenStack-dev mailing list