[arc-discuss] [ogb-discuss] Defining OpenSolaris Consolidations

[Stephen Hahn]

* Keith M Wesolowski <keith.wesolowski at Sun.COM> [2007-08-07 10:57]:
> On Tue, Aug 07, 2007 at 09:43:05AM -0700, Stephen Hahn wrote:
> 
> > > An OpenSolaris Consolidation is a Project that has been sponsored by
> > > an OpenSolaris Community, for which the Contributors have agreed to
> > > follow the policies set forth by the OpenSolaris W-Team, and has been
> > > accepted by the OpenSolaris W-Team.
> > 
> >   Can any Project become a Consolidation, or must this desire be
> >   declared up front by a candidate?  
> 
> I don't think this needs to be declared up front, but a project must
> (of technical necessity) declare whether or not it will target
> integration into a consolidation.  Only a standalone project is
> eligible to become a consolidation; it has the properties of a
> consolidation:
> 
> 1. Its contents do not duplicate those in any other consolidation, and
> 
> 2. It is effectively parentless(*).
> 
> The first of these is Alan's requirement (6); the second is an
> implementation detail.  I suppose a project could elect at any time to
> switch between these two states, so long as it has not been granted
> consolidation status.
> 
> (*) The parent of a consolidation is a previous incarnation of itself.
 
  I'm tired, or something, because I can't figure out what would make a
  project standalone or not standalone *and* I can't understand how
  parentless applies to a new consolidation.  Maybe there are better
  terms available?

> >   Can Consolidation status be removed?  Which parties can initiate this
> >   removal?
> 
> If a project were no longer endorsed by any Community Group, it would
> be disbanded entirely.  Otherwise, it seems that the W-team has this
> authority based on Alan's description.
> 
> > >   4) That all original contributions to the Consolidation be made under a
> > >      Contributor Agreement assigning joint rights to the OpenSolaris IP
> > >      Steward designated by the OGB.   The OGB's initial designation of
> > >      IP Steward is Sun Microsystems, under the terms of the current
> > >      Sun Contributor's Agreement.   (Requirements for acceptable licenses
> > >      are already listed in os_dev_process.)
> > 
> >   I don't think, under Charter, the OGB designates the Steward, but I
> >   agree with the Contributor Agreement requirement.
> 
> Why not?  Article II requires only that the software we produce be
> "licensed to the public free of charge under one or more open source
> licenses approved by the Open Source Initiative".  It also gives the
> OpenSolaris Community the "authority and responsibility for all
> decisions pertaining to the OpenSolaris software."  If anyone can
> impose such a requirement, the OGB can.  The real question is why we
> would want to.
> 
> It is not "inconsistent with applicable laws and regulations" to do
> away with this requirement, and I see nothing at all allowing Sun to
> impose it.  If we keep it, it should have an expiration date - 1 year
> might be reasonable - and should be expressed as a gift to Sun given
> freely in appreciation and reciprocation.  But I'd have to wonder
> whether doing so is inconsistent with another of our Article II
> obligations, that of "fostering the evolution and adoption of the
> OpenSolaris code base".  At this point, I simply don't see how the
> existing requirement helps anyone other than Sun (the corporation, not
> any of its employees, whose work is owned by Sun regardless).

  Second paragraph of Article 6 of the Charter, I would have thought.
  It might be possible to accept contributions from any of multiple
  Stewards, I suppose, although that could ultimately lead somewhere
  messy.

> > > Once at least 3 OpenSolaris Consolidations have been designated, a distro
> > > claiming to be "based on OpenSolaris" should have a significant portion of
> > > at least one of the OpenSolaris Consolidations.   A distro claiming to be
> > > "an OpenSolaris Reference distro" must contain the entirety of all current
> > > OpenSolaris Consolidations at the time of the feature freeze of its release.
> > 
> >   I think there are problems with "contain the entirety", but I think
> >   exploring these OpenSolaris-ness tests is worthwhile.
> 
> Yes, but it's also problematic to suggest that a "reference
> distribution" could include anything *more* than what's in the
> consolidations - otherwise, those contents may be something to which
> someone cannot in fact refer, defeating the entire purpose.  This
> definitely requires more thought.

  I confess to only being interested in arguing for a "defined set
  potentially smaller than the entirety" case.

  - Stephen

-- 
sch at sun.com  http://blogs.sun.com/sch/