Peter Tribble wrote:
> On Thu, Jul 9, 2009 at 7:07 PM, Garrett D'Amore<Garrett.Damore at sun.com>
> wrote:
>
>> Now, as an aside, I believe that the business teams that have members on ARC
>> benefit greatly by it. The ARC members are able to help shepherd that teams
>> projects, and mentor the rest of the staff, so that when projects come to
>> ARC they are more fully formed, and streamlined in a way that minimizes
>> contention at ARC.
>>
>
> Should all cases have some level of review before getting to ARC? That weeds
> out some obvious shortcomings, even if the review doesn't include someone
> with ARC experience.
>
Automatically an ARC sponsor is supposed to review the case, so yes,
there should be some initial review. But some licensees are better at
performing this review than others. In general licensees that have
actual ARC experience are better (IMO) as sponsors than those that have
never interned before.
> I know that a number of webstack cases were open to review by the community,
> which helped flush out some of the details. Waiting for the ARC case to ask
> basic questions isn't optimal; having the ARC case be the first you've heard
> of
> the existence of a project is worse.
>
Yes, that's true!
> Which opens up the question: how can the OpenSolaris community be effectively
> engaged to improve the review process? There is significant expertise there,
> although many of us cannot commit to schedules. Still, having (for example) a
> case having had a pre-review by an OpenSolaris community group would ensure
> that it was at least basically sound, leaving the main ARC review to
> concentrate
> on how it fitted in the wider context.
>
The individual community groups need to resolve this. Many of them
already do that... the Networking group is actually really good at
this. Some other groups also are pretty good at it.
Most of the issues we have with lightly reviewed or unreviewed cases are
actually those "familiarity" cases that come from folks without any
engagement with a larger community group. (IMO, many of these cases
originated within Sun -- there was at one point a belief that some
engineers had that they were going to be evaluated based on how many
FOSS projects they got integrated into Nevada. Fortunately, the
pressure behind that mentality seems to have eased somewhat. Perhaps
the failure of a few such cases to pass even basic ARC review has caused
a few people to rethink that particular strategy somewhat. At least
most recently the cases that seem to have come before ARC for FOSS stuff
seem to have at least a modicum of sanity behind them.)
- Garrett
