On Fri, Sep 30, 2011 at 10:33 AM, Rory O'Farrell <[email protected]> wrote: > On Fri, 30 Sep 2011 10:15:52 -0400 > Rob Weir <[email protected]> wrote: > >><snip> > >> By my count we have 72 committers right now, almost all of them also >> PPMC members. With the new IBMers coming on board, as well as >> possibly forum admins/moderators/volunteers (at least according to one >> draft proposal), we could shortly have 120+ committers/ppmc members. >> >> This is not just quantitatively different. This is qualitatively >> different. It is no longer a committee. It doesn't work like a >> committee. It doesn't think like a committee. It is not necessarily >> a bad thing, but it is qualitatively a different thing. It is more >> like a congress, where factions form and individual voices are less >> heard. The voice of reason is less often heard in a "committee" of >> 120 people. It doesn't cut through the noise. >> > > Without wishing to engage in this discussion in any depth, I point out that > the accepted wisdom is that the optimum size of a decision making body is > between 6 and 12; our ancestors felt twelve to be optimum (hence the > traditional jury size). My own experience is that about seven to nine is > good, if one accepts a majority vote rather than unanimity. >
I think the potential problem that Ross points out is that Apache PMCs select their own members. They are not representatives in a formal sense, they are not selected by the broader membership. So that could hypothetically encourage stagnation of ideas, etc. There is a reason why close inbreeding is discouraged in most cultures. One way to avoid that is to increase the gene pool by having everyone be in the PMC. But that has the potential to degrade the effectiveness of the PMC's decision making progress. The ideal, IMHO, is to have a PMC that is right-sized, and whose members have the confidence of the overall project. A size large enough to fulfill its responsibilities, but small enough so that every member of the PMC can be fully informed on the issues it is deciding on. Gaining the the confidence of the overall project is the tricky part. One way is to have the PMC be elected by the general project members. But that isn't how Apache works. The PMC elects its own members, and debates these votes in private. In theory, this is because we're a meritocracy, not a democracy. Maybe one way to remedy this is to have PMC votes and discussions take place on the public list, with prior permission of the candidate? You could argue that since it is a community-wide role, the input of the entire community would be valued in making that decision. -Rob > -- > Rory O'Farrell <[email protected]> >
