[openstack-dev] The future of Incubation and Core

Russell Bryant rbryant at redhat.com
Fri Nov 9 17:50:05 UTC 2012


On 11/09/2012 12:13 PM, Jay Pipes wrote:
> Well said.
> 
> After reading through this ML thread multiple times and really trying to
> give all viewpoints a thorough amount of time to sink into my brain,
> I've come to the following conclusions:
> 
> 1) The idea of "core" vs. "supported"/"ecosystem"/"related" is really
> NOT all that important to users (whether users are service
> providers/operators or developers or IT departments or universities or
> anyone). The distinction is immaterial to users of OpenStack in the same
> way that a distinction between an engine and a steering wheel are of no
> consequence to a driver of a car -- while the car may not run without
> the engine, it's pretty pointless to the driver if he can't steer the car.
> 
> 2) The idea of "core" vs. "supported"/"ecosystem"/"related" DOES matter
> to product and business development folks since the Board of Directors
> uses the concept of "core" vs anything else in its evaluation of
> trademark enforcement.
> 
> 3) The idea of coordination -- in releases, in Development/CI/QA
> tooling, in community management, and in the approach that APIs and
> implementations are transparently developed -- is the REAL topic of
> discussion for the Technical Committee, not what is "core" and what isn't.
> 
> 4) Inclusiveness matters. Exclusive communities tend to stagnate whilst
> inclusive communities tend to grow -- with growing pains, but they grow.
> OpenStack should be a growing community, not a shrinking or exclusive one.
> 
> In conclusion, I came into the Technical Committee meeting this week
> with a set of opinions that I have pretty much maintained for over a
> year now: that "core" should mean infrastructure-only projects and that
> we should only incubate projects that are considered
> infrastructure/foundational.
> 
> After the meeting and this discussion, I've changed my mind and I no
> longer think it is useful for the Technical Committee and the technical
> community to distinguish between "core" and "supported". I think that
> incubation should refer to the following:
> 
>  A joint commitment between the proposing project and the existing
> OpenStack developers, CI, Community/Release Management and QA teams to:
> 
>  * Adopt OpenStack common practices and tools used in developing and
> testing the project
>  * Adopt OpenStack common practices for tracking documentaion, bugs and
> a roadmap
>  * Agree to coordinate releases of the project **at least as often as
> the OpenStack 6-month release cycle**, though the project is able to
> release more often as desired.
>  * Work with other OpenStack projects to reduce integration point
> friction and better consume/utilize existing OpenStack APIs
>  * Use a single 6-month release cycle to "blend in" with the
> above-mentioned tools and practices
> 
> A proposal by a project to "incubate" would merely then be a simple
> agreement of best faith to follow the above things and nothing more than
> that. At the end of the six month incubation period, the Technical
> Committee would be asked to decide whether they believe the project
> tried in best faith to adhere to the above points, and if not, decide
> whether to rule the project as "not OpenStack", whether to include the
> project in OpenStack and/or to give additional incubation time and some
> suggestions for areas of improvement.

Very well written, Jay.  +1 to all of that.

Though I'm not entirely clear on your opinion on the future of "core".
You acknowledge that it has importance to the board for the purposes of
trademark enforcement.  Is that something that we (the TC) should just
push off to the board to handle?  Or should the TC still be involved in
that?  If so, to what degree?

-- 
Russell Bryant



More information about the OpenStack-dev mailing list