On Wed, Jun 06, 2007 at 11:17:55AM -0600, Mark J. Nelson wrote: > >I'm not arguing that any project team should do this -- only that noone > >can stop them from trying, EXCEPT that the OpenSolaris ARC must have the > >authority to deny/issue TCRs. > > If you do not allow closed review, you are practically begging for teams > to circumvent the process in order to keep information private. > > Essentially, the teams in question have two legitimate goals: get > effective, timely review; and keep proprietary information private. If > your process makes those goals mutually exclusive, and the second is > mandated by contract (or patent law or whatever), then your process is > effectively denying the first goal.
Well, project teams today do seek informal advice from ARC members before they reach the ARC. We might formalize that, but at the end of the day (i.e., when we reach a steady state on open process) the bottom line could be that there has to be an open case, and it has to be de novo (though archives from closed review should be made available to help speed up the open case review). Or mgmt could tell us to build in an escape clause for secret cases (have they?). > >But I don't see how OpenSolaris can prevent project teams from believing > >that they can take this tack -- OpenSolaris can only refuse to > >rubber-stamp (and it should). > > You can prevent successful end runs by having, and appropriately > exercising, authority. Isn't this thread about preventing them? I guess it isn't, actually, since it doesn't the proposal doesn't tackle the c-team aspect of the process :) > You can minimize attempted end runs by crafting a process that enables > teams to achieve all of their goals, and then using that process to > facilitate that achievement. It doesn't happen overnight; teams will > initially be suspicious of your motives, and will not understand the value > you provide. Yup. The whole changing engines in mid-air thing. > >>>... But Sun, in particular, has an OpenSolaris-derivative product > >>>(Solaris itself), and I think it behooves Sun not to run closed cases > >>>that integrate into the closed portion of Solaris just to avoid public > >>>exposure. > >> > >>Really? And management agrees? > > > >I really think that; I don't know if management agrees. > > You cannot address this at the process end. The process must facilitate > the needs of the community. In this case, there are folks inside Sun (and > eventually XXX-corp) who need to avoid public exposure. > > You can only fix that at the source: the real or perceived need to keep > information private. Until you fix that, a process that prevents it will > not succeed. OK, that makes sense. We can probably fix "perceived need[s] to keep information private." But what about real ones? See below. > >>>Sun projects that wish to stay out of the public view until ready should > >>>just fine tune their timing in view of the need for running ARC cases in > >>>the open. > >> > >>That, in my opinion, is just plain broken. > > > >Huh? What's broken? Are you saying that we need closed ARC cases? > > Yes. > > It's broken for the reasons I described above. You're talking about > giving a project team an ultimatum: either break your legally binding > contract, or forego architectural review. I don't think it's unreasonable to ask for open review of things that want to integrate into OpenSolaris. Solaris != OpenSolaris, so projects that require secrecy should integrate into closed Solaris if they wish to integrate into Solaris before running an open ARC case. The problem will then be that some day they'll want to integrate into OpenSolaris and offer their existing Solaris architecture as a fait accompli. I don't have answers to all problems :/ > That's a crappy, broken process. Yet you want to avoid the need for closed cases (you just don't want to address that from the process side of things right now). I agree. I'm not arguing that the moment this proposal is adopted then all Sun internal projects that intend to eventually integrate into OpenSolaris must obtain open ARC review. Just that as long as we have mgmt telling us to oepn Solairs then we have to aim for openness, and where that causes conflicts with Sun business then mgmt has to be made aware. I'd be happy with us saying that there will always be room for Sun-internal projects to integrate into OpenSolaris without open ARC review. As long as that's clearly stated in the OpenSolaris ARC charter so the community that we're trying to build understands it. If the community votes with its feet we can always go back to trying to be more open. > >The decision to open Solaris came from the top, from Sun's CEO. Ergo > >Sun management adopted an openness policy. Perhaps all the consequences > >haven't been worked out and perhaps Sun management may change its view > >on openness, but for now the dictum we're working on is that we're > >trying to build an open community around OpenSolaris -- the dictum is > >openness. > > If the problem is the corporate need to keep information private, then the > solution lies with the corporation, not the process. Yes, but we're finding the problems in the process of working out the process. > I think it's disingenuous to deny that need, and I think that trying to > force change by implementing an incompatible process will result a process > that is not followed. I don't deny that need. My comment on forcing mgmt was in response to Jim's comment that "[w]hile we're forcing Sun management into openness..." That is, I'm not pretending to be able to force mgmt to do anything; but if mgmt sets a direction and we run into problems with it then one way or another mgmt will have to deal with them. (To get down to editing nits: I should not have said "and live with" -- Sun mgmt gets to modify the openness policy until it suits Sun.) Nico --
