[arc-discuss] OpenSolaris, ARCs and the future.

[Darren Reed]

At the bottom is a pointer to the type of project formation
discussion that I think OpenSolaris should be aiming to reach.
This style of discussion fills in the "ARC" gap.  Ultimately,
I'd like to think that OpenSolaris can move away from having a
centralised ARC process to each community being its own ARC.
I should add that not all project proposals at NetBSD go forward
as smoothly as the one below as the membership is all humans
and thus discussion threads diverage and digress.

Some assorted ideas/thoughts...
Today the operation of PSARC centers around the members
who are supposedly from a cross section of Sun but in reality
it has a large percentage of its members from a much smaller
cross section.  There are a couple of problems here:
1) it's unreasonable to expect ARC members to be experts on
   the architecture of all aspects of OpenSolaris and as the
   bredth of the membership shrinks, there is increasing risk
   of this happening;
2) as a reuslt of (1) it becomes more likely that projects are
   less well reviewed than they should be.

A couple of caveats:
3) if a project's ARC review happens within its own community,
  areas where it touches on the architecture of other components
  of OpenSolaris may not be as well reviewed;
4) a community is may be more inclined to approve a project with
   less than desirable architecture,, especially given the almost
   "automatic" "+1" replies from people to keep things moving.

But I'm not sure how to solve our problem without creating more process.

Something that I'd like to put out there for people to think about is
changing the membership of the ARCs to create a level playing
field, where there are only "licencess" (no members, perhaps
also no interns but i'm not convinced) that all have the same
responsibility and authority.  That is, a project could be "derailed"
by "anyone" if they were able to justify the technical merit.  Thus
if someone comes along with a project that uses SMF in some borked
fashion, there is more scope for someone from the SMF community to
speak up and see that justice is done.

Thoughts?

...I'll admit that where I've used "project" above refers soley to a
project that involves hacking on code for OpenSolaris and not other
random tasks, such as advocacy, etc, and that where I've referenced
"community", I again refer to those involved with hacking on code.

http://kerneltrap.org/mailarchive/netbsd-tech-kern/2007/10/5/330511

Darren
p.s. is it appropriate to also send this to psarc@ or psarc-ext@?