Open Stack

Gabriel Hurley wrote:
> As to the Program proliferation, it just struck me as "Queue" being an exceptionally specific program compared to many of the existing ones and wondered if we'd accomplished anything by changing from Projects to Programs. One of the goals of that was to try and slow the explosive increase in management overhead, right? If no one else cares, I don't either. I was on the side of "more projects isn't an issue" when that debate happened before; I just wanted to raise the issue for discussion because previously folks seemed to care a lot.

The main goal of programs was to acknowledge that there were teams
working on stuff that may not result in a release-integrated
deliverable, and that those teams were as important to OpenStack as
anything else.

The second goal of programs was to add a layer of abstraction between
projects (repositories) and teams that handle them, to better account
for teams producing multiple things.

I don't think introducing programs makes management overhead any worse
or better. It used to be an issue to add projects when this was directly
linked to TC membership, but now that this was decoupled, I don't think
it's that much of an issue. I see a problem (and management overhead,
actually) with creating an artificial structure to serve as an umbrella
for completely-separate teams working on completely-separate projects,
though.

That doesn't mean those projects could not live under the same program,
but they would first have to work together and be, in effect, one team.
I would welcome them to merge, but would not externally impose it. In a
highly-organic system like OpenStack, structure should reflect how
people work, not the other way around.

-- 
Thierry Carrez (ttx)