[arc-discuss] Requesting PSARC/2007/124 details

[James Carlson]

Alan Coopersmith writes:
> James Carlson wrote:
> > What's the perceived need or underlying threat here? 
> 
> The need is figuring out how a Sun-controlled body like SAC/ARC fits into
> a world where we are participants in many projects but controlling none of them.
> 
> Why should Sun's SAC have special powers over OpenSolaris that no other
> participant in OpenSolaris has?

I don't think it should.  I'd go as far as saying that there's a
structural problem in just talking about "Sun's SAC."

Besides sponsoring the ARC and the somewhat more obscure BUSGs and
RBs, one of the important roles that SAC plays is as part of the
appeals process.  I believe we _must_ have an open appeals process to
go with our open architectural review process.

Secret appeals are just a non-starter, so this process needs to be
rewritten.

The rewrite will undoubtedly be difficult, as one of the things the
SAC was chartered to deal with in those appeals is the occasional
conflict between business and architecture.  "Business" now means the
goals and strategies of all of the distributions, not just Sun's.

Also, the process currently makes quite a few references to Sun's
internal organizational structure and to our seniority system.
Neither has any place in OpenSolaris.

There are miles to go.

>   If another major vendor with internal
> architectural review processes join the community, would they get similar
> control?

I think all of the distributions should have equal say at that level.

But for normal day-to-day activity, an open ARC solves the problem, as
we normally (at least 99% of the time) resolve things by discussing
them, not by escalating.  (And the times we do escalate, the answers
are almost always bad ...)

> If Sun is in no way special, then why would Sun's ARC's tie all these
> unrelated communities together?   Why should the OpenJDK architectural
> review be tied more to OpenSolaris than to Linux or BSD?   Or are we
> opening up our ARC processes to tie together any open source projects
> that want to interoperate?    Will we create OpenSAC that may have
> participating projects from any community, regardless of Sun participation?
> Sort of a loose affiliation of projects like Apache Foundation or
> FreeDesktop.org?

It's an interesting question, but not one I think needs to be answered
right away.

The key issue here is that there are multiple cooks in the kitchen.
In the past, we've been able to deal with this by drawing lines on a
case-by-case basis.  I don't see why that will fail to work tomorrow.

So, yes, I do think that some projects, by their very nature, are
going to have to be reviewed by the ARC as well as by other bodies --
because those projects touch things that have effects across multiple
components.

Simply saying that JDK bits always go to OpenJDK and never confer with
either Linux, BSD, or Solaris sounds to me like a mistake.  That's
likely true for most cases, but just as likely not true for some
cases.

-- 
James Carlson, Solaris Networking              <james.d.carlson at sun.com>
Sun Microsystems / 1 Network Drive         71.232W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442 1677