On Thu, Jan 24, 2008 at 01:11:46PM -0800, Roy T. Fielding wrote: > is different from the outside world. I would have preferred to > eliminate > all of the community discuss lists (meaningless distractions) and > centered everything around projects (just like Apache).
So would I. But there's a place for SIGs, too, which is what the CGs have effectively become. What Apache calls a project is really a product. Products, SIGs, and projects are all different things. CGs have attributes that seem reasonable for products. They are worse than useless for either of the others, and worse still for the various other purposes to which they're being put today. > > The problems noted above with the LDOM's effort are almost > > identical in > > the case of distributions. Should Core Contribs of a Distribution CG > > make decisions about Belenix, Nexenta, Schilix and Indiana? I don't > > think so. Never-the-less, there should be a grouping which I call a > > "meta-community" or "Consortium" which can provide a loose social > > grouping of CG's to facilitate collaboration and coordination. > > They should be given the choice to do so. This is essentially what > ARC is imposing right now -- an architecture for OpenSolaris. Doesn't > it make more sense to at least know what all distros are willing to > agree to implement? Yes, but it depends what you intend. Suggesting common architecture or functionality within the "Distributors' Guild" seems fine. Advising the ARC of those common desires also seems fine, even beneficial. But one may join or not as one chooses. The ARC is different; it's often impossible to avoid the consequences of an ARC decision without taking a complete fork of the source base and maintaining it in perpetuity. So a Distributions SIG (call it a CG if you like; the name does not change what it really is) cannot replace the ARC because it does not and can not hope to represent all consumers of the technology. > By all means, spend the next three years on more meta-discussions > instead of actually using what has already been agreed to. I think > you should try using the powers that have been given to the OGB. What we have will not work for us. Period. I'd rather spend the next three weeks on meta-discussions and then replace most of the constitution we have today with something that could possibly work. Then we can stop wasting time on this crap and go back to creating software. Sure, we could terminate most of the CGs that don't make sense, but what good would that do? We don't have the structures we need to represent the things we care about, so there's nothing better with which we could replace them. The OGB's powers cannot solve this problem unless we consider wholesale fabrication of stuff not even imagined in the constitution to be within those powers. I'd much rather just replace 2/3 of the document with the words "My development process, let me show you it: http://www.opensolaris.org/os/community/on/os_dev_process/";. That step would more than make up for in correctness and sanity what it would cost us in completeness, precision, and ideological sisterhood with other successful organisations. I will say that creating a Distributions CG is a definite step in the wrong direction. At best, it's another SIG: possibly of some value to discuss some of the meta-issues that others have raised about distributions in general, but not something that would take us any closer to a use of CGs that is actually consistent with their definition. Worse, it would seem to actively preclude representing individual distributions as CGs - the one possible use for which they seem a good fit. In one small way, then - denying Mr. Walker's request - the OGB may hope to use its powers to at least preserve the possibility of something that makes sense. > How does changing the names in the constitution change that? If you It doesn't. The names aren't the problem; the definitions and relationships are. sed -e s/Community Group/Project/g doesn't make it a better structure for a Consolidation, for the ARC, or for a group of people who want to work on DTrace. Nor does it make any of those things into a bunch of hackers spewing out a piece of shrink-wrapped software, which is what Apache and everyone else seem to call a "Project". > More importantly, I don't see Sun giving up control over the gates. Depends what you mean by control. I do believe that there will come a time at which anyone will be able to access them directly. I also believe there will come a time when it will be possible for legitimate modifications to them to occur without the involvement of any Sun employee. What I do not know is when either of those milestones will be reached. If your theory is that the OGB should simply determine that the gates on the SWAN are not the real OpenSolaris gates any longer, I'd like to know with what you would have us replace them. -- Keith M Wesolowski "Sir, we're surrounded!" Fishworks "Excellent; we can attack in any direction!"
