On Fri, 28 Dec 2007, Al Hopper wrote: > On Thu, 27 Dec 2007, Alan DuBoff wrote: > >> On Wed, 26 Dec 2007, Roy T. Fielding wrote: >> >>> The likelihood of a sound architectural decision being rejected >>> by the core contributors (almost all of whom are Sun engineers) >>> is quite small. >> >> That's all arbitrary. The fact is that nothing substantial has been >> proposed by anyone, all of the meat has come from Sun. Yes, it's a good >> thing that Sun's engineers are still working on it, because the external >> community hasn't really come up with any type of sound architecture >> decision, AFAIK. > > Not true. Case in point: the ksh93 integration project.
Al, When I read Roy's comment, I took it for face value, "sound architectural decision". Don't get me wrong, it was very valuable for the community, and Roland specific did much work to get it putback. Knowing exactly what is involved in a putback, my hat is off to anyone doing so. But do you consider ksh93 to be a "sound architectural decision"? It was integrating a piece of open source software from my view, and replacing a component that was already in place. > This project > touches on so many areas of the OS ecosystem, that it is probably one of the > most difficult/complex integrations done to date - and it was done by a > community contributor (the unstoppable) Roland Mainz. In fact, it was the > contentious (technical) infighting within Sun (engineering) that prevented it > being done for many years. Agree with you on that, but it has nothing to do with the architecture of (Open)Solaris. Even so, a community member can't just say I'm going to putback ksh93 without getting the proper approval and complying with the formal process put in place. The community member must have a sponsor to do so, and in fact, the community has little bearing on what is accepted/approved by any of the process that is entirely inside the SWAN. The bottom line is that the community can't do $#!T as the system is today, without having someone to wipe their @$$ for them. How much influence did the CAB have on OpenSolaris? How much did the CAB dictate on what was putback? How much power does the OGB have today? What do they actually approve or not approve in what is putback? I would be first to agree that it is changing and moving in that direction, but don't jest me with pipe dreams until they come true. > But, in most, as in 98% (Wild Ass Guess), of the cases, a community > putback can't be blocked if it is technically sound and meets the tough > reliability/quality (etc) standards imposed by the ARC. BFD, everytrhing looks good on paper. The bottom line is that the community can't even putback unless they have someone help them. Nobody, not even a core contributer can putback unless they're on SWAN, and can get the approval for their putback. Technically, you only need to get approval from the gate keeper and tech lead, but there's a lot of eyeballs watching the process so there's many more players to get by for any putback to happen. > If this is not the case, or, if this is not the *perception* of all the > current, or potential, community contributors, then the (OpenSolaris) > Project is doomed to failure. I think we're moving in that direction, but aren't there today, and we aren't going to be there tomorrow, or even next week, it will take a little time. This is not bad, I'm just pointing out that you can say that the CAB/OGB has the power to control things that happen, but in reality they don't have enough power to really blow their own nose, without having to ask for a tissue. > From my personal observations of the current process and procedures, > which may possibly be flawed, the current system appears to be working, > as exemplified by the ksh93 integration project, with the obvious > current exception of the mfi project, for which, AFAIK, no technical > deficiencies have been identified which would/could prevent integration. I'd say they should get a sponsor and "go for it", sounds pretty easy. Should be slam dunk to get the mfi putback. > It is vitally important, IMHO, that the mfi project be integrated - or > valid technical obstacle(s) be identified, by reasonable community > involved consensus, which prevents it being integrated. I tried to explain some of what I am aware of. This is a case that will not be so cut and dried, since the megasas that is in development, and the fact that the mfi doesn't have full support for the same devices, AFAIK. After break I'll try to get my manager to give some type of statement here on opensolaris for where the LSI work stands and/or what the licensing *might* be, because until it is final it could change. Rich Lowe had an interesting thought on trying to leverage the IHV gate to handle this type of situation where a driver that might not be open source could be in the IHV gate, and an open driver in the OpenSolaris gate, however this is all speculation as I'm not even clear if the megasas is open source or not, and I don't think anyone else on this thread so far knows either (at least hasn't mentioned it if they do;-). > Remember that perception is in the eye of the beholder. If the > community does not share the *perception* that all but technically > suspect contributions/changes can be successfully integrated, then the > OpenSolaris project will not gain the mindshare it deserves and the > technically proficient contributors will contribute their expertise > elsewhere. I think we're moving in that direction, but we're not there today, and any community member trying to kid themself into believing that the process does in fact allow this to happen is very naive. It is entirely possible that I do not understand the system and/or how the putback is done, and that the community has a lot more power than I can see myself. From my view it's all smoke and mirrors as it is today. The only difference is that Sun tosses a tarball over the firewall each night from the build. The community has very little influence on what happens to get the tarball, they only get consideration from the current/existing process that is in place, but that is not even mandatory. IOW, I think Sun has changed it's thinking some and is moving outside the firewall, just ain't there today, so I have a hard time buying into any argument of how the community has any power whatsoever on the putback process until that happens. I haven't seen a single putback that the CAB/OGB has ushered through the existing process. -- Alan DuBoff - Solaris x86 IHV/OEM Group _______________________________________________ driver-discuss mailing list driver-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/driver-discuss
