* Eric Saxe <Eric.Saxe at Sun.COM> [2008-05-15 01:14]: > Stephen Hahn wrote: > >* Randy Fishel <randy.fishel at sun.com> [2008-05-14 21:06]: > > > >>On Wed, 14 May 2008, Alan Coopersmith wrote: > >> > >>>Randy Fishel wrote: > >>> > >>>>I would like to propose starting a Power Management Community. > >>>>The intention is to coalesce fragmented power management > >>>>discussion and work into a single community. > >>>> > >>>> The proposal can be found here: > >>>> > >>>>http://www.genunix.org/wiki/index.php/Power_Management_Community_Group_Proposal_2008 > >>>> > >>>The OGB discussed this Monday, and had 3 questions before approving: > >>> > >>>1) Why a community instead of a project? > >>> > >>Primarily, because it would never be a single project. There is a > >>larger desire to coordinate power management work, instead of having > >>bits (possibly conflicting bits) scattered all over the place. And it > >>is also difficult for developers/users to know where they should go to > >>address their specific problems or needs. The community could well be > >>considered an umbrella, or meta-project, as one goal is the creation > >>of projects to solve specific needs, but that grouping doesn't exist > >>(yet). > >> > > > > One of the problems I see with this proposal (and the Emancipation CG > > proposal, for that matter) is that the larger scale tradeoff > > discussions aren't going to happen in the proposed CG, but in the CGs > > that hold the responsibilities for the main source tree for each > > proposed change. That is, I'm sure that Power Management participants > > might all agree that a particular change is great for Power > > Management, but that a person more interested in performance or > > availability might disagree. That discussion happens within ON. > > > I'm not really sure that there exists a way to properly organize the > "larger" OpenSolaris community into smaller groups that could eliminate > this problem. I can see that any of the existing communities today could > sponsor project work that winds up taking too myopic of an approach.
Actually, for our technical efforts, it seems pretty clear that the large bodies of code and contributions to that code are primary organizing factors. If anything, ON has been hamstrung as a CG by the fact that many of its major areas have been carved out into CGs of their own. I'd rather get Simon's proposal (or reread one of Keith's similar ideas) into place before going ahead. > Those projects *should* be diligent enough to solicit design and review > feedback in the broader forums...and in case they don't...as these > projects seek to contribute their code to the repositories owned by the > larger CGs, it should arise though the review process (e.g. CRT), that > due diligence hasn't been done. This hope would be sufficient if people never disagreed. The whole point of the governance model is how we handle disputes: I would prefer that a dispute involving changes to ON be handled by voting among the core contributors to ON, and not by ON and some other (presumably SIG-style) CG. > The amount of power management work that could (and should) happen > through over OS.o is too large and diverse to properly fit under any > existing project (like Tesla)...and as a leader of that project, I would > very much appreciate the opportunity to participate in a larger forum > that would facilitate collaborating with the other power management > related efforts (like storage PM, memory PM, device PM, efficient > software development practices, etc) that can and should be working and > participating in the open. It's pretty clear that the problems relating > to power management in particular cut well across the system stack, so a > forum that helps to facilitate collaboration in this area would be helpful. You keep using the term "forum". You can't mean a mailing list and web pages, because projects have those facilities as well. I can only assume you are really hoping for "exclusive or sole forum". Of course, a CG doesn't get you exclusivity--by design, since we've already seen that at times, due to lack of interest, CGs die out, and their goals get partially assumed by individuals in other CGs. I think that, if exclusivity or highlighting is what you're after, then we should wait for the Board to come up with some kind of special interest group tag for projects or mail aliases. I'm a strong -1 on yet another source of Core Contributor grants for work that will end up in ON. I'm also a -1 on an Emancipation CG for these reasons (as well as not understanding how to evaluate contributions to Emancipation work that don't ever end up in a mainline source tree or a distro). - Stephen -- sch at sun.com http://blogs.sun.com/sch/
