John Plocher wrote:
> Bill Sommerfeld wrote:
>> On Thu, 2007-04-12 at 14:47 -0700, John Plocher wrote:
>>> This actually may be a good thing, since in the long run (post-
>>> PSARC-transition...) the ARC community should IMHO explicitly be a
>>> derived one - the Core Contributers from all the other communities
>>> should be defined to be the ARC Community.
>>
>> so I don't see why this should be the case.
>
> Ask yourself "who" should be reviewing changes.
>
> The answer that I come up with, in general, is
>
> We all are responsible for managing the evolution
> of the architecture of the things we build. The
> groups that are impacted by changes need to play
> a role in their review. The scope of "who" is
> related to the scope of the proposed change: is
> it local, within a component, or exposed for others
> to use?
>
> A simple bugfix is usually reviewed by a responsible
> engineer and her immediate team members. In OS.o land,
> this maps to the various contributers of a project.
>
> If the bug/rfe impacts things outside a project, but
> within a consolidation/component, the scope of reviewers
> widens to include other members of the component's
> development team. In OS.o, the idea of a component
> is mixed in with the idea of a more amorphous Community,
> so it is difficult to say who should be involved here.
>
> If a proposed change exposes or changes externally visible
> artifacts, then the developers of the other components
> that depend on it get involved, because this sort of
> change impacts the systems that OS.o part of. In the
> OpenSolaris environment, the core contributers of the
> various communities are the people that need to get
> involved with this level of review.
>
>
> If not these people, then who should be the decision
> makers? Some group of engineers who are not involved
> with OpenSolaris? Some group who are not architects in
> their own part of the community?
I think you may be misunderstanding Bill's response.
I think we can form a reasonable ARC community from senior
members/leadership who _are_ participants in OpenSolaris.
I do not believe it is sound to believe that every OpenSolaris community
has leadership that have the technical judgement to guide OpenSolaris.
I do believe that we should try harder to get people who are currently
employers of a major OpenSolaris contributor to also become part of the
_public_ process. That means that some of those people should probably
be participating here.
Some of them already do... we even have ARC representation on the OGB.
I do not think it is a good idea for folks in the user-groups community
to be weighing in on kernel land APIs, though. (At least not if their
only qualification is that they run a user group somewhere.)
All this goes to say, I don't think you can automate the ARC selection
process. I'd like to think that the ARC is a kind of cabal, where new
members are selected by the existing members. I'd be willing to allow
that the initial set of ARC members should be selected probably by OGB,
and that it would be a good idea to start by extending an invitation to
the existing Sun-internal ARCs.
I'd also be willing to allow that OGB could have an some kind of
oversight of ARC selection in the future... perhaps by having the
ability to nominate members on its own, or by having some kind of
confirmation from OGB... although, really, OGB isn't supposed to be
meddling in technical affairs.
And that's an important point. As I understand it, ARCs are _technical_
entities making _technical_ recommendations/judgements... they don't do
marketing, they don't do docs, and they try to avoid politics. Please
keep that in mind.
-- Garrett
