> On Apr 23, 2017, at 11:09 AM, Shane Curcuru <a...@shanecurcuru.org> wrote: > > Pat Ferrel wrote on 4/22/17 11:46 AM: >> Probably the wrong place for this but… >> >> What do people think about a governance change for approving releases >> through the IPMC to wit: >> >> A week of no vote activity over the release proposal of a podling >> should be considered a passing vote. In other words the IPMC is to >> become a vetoing group. > > No, because as noted in this thread the board and Legal Affairs > Committee already have a policy for Apache software releases: > > https://www.apache.org/legal/release-policy.html#policy > > This is the spot where all the little legal bits of having a corporation > intersects with the Incubator's ability to mentor new podlings in how to > effectively run projects in the Apache Way. However that legal policy > in particular is not going to change IMO. > > Personally, I support any proposals that help ensure podling mentors are > more actively engaged, and can better serve as the formal IPMC votes for > podling releases. Great to see some serious discussion about that in > this thread and the upcoming IPMC chair election. The incubator has > been getting better organized as a whole over the past few years, and > keeps improving, which is great to see. > >> I propose this for 2 reasons: 1) lack of votes or attention from the >> IPMC seems all too prevalent and puts an undo drag on the energy and >> velocity of community involvement. 2) the release has already been >> voted on and checked by at least 3 PMC members of the podling (which >> has ASF mentors in most all cases) so a veto role by the IPMC seems >> to have minimum danger to the ASF system of checks and balances.> > > Yes, but the whole point of Incubation is that a podling is not yet a > project. Only Apache PMCs have the authority to make a release, thus > only IPMC votes count. On one hand this must feel incredibly annoying. > On the other hand, it is the #1 reason we have the ASF as a corporation > - to prevent the individual release manager from potentially getting > sued *personally* for problems with the release. > > The fact that we have these documented policies, and that the board and > IPMC enforces them means that legally, releases are acts of the > Foundation. Thus if anyone ever were to sue, they'd sue the Foundation > (which has insurance and legal counsel) and not individual committers. > > -- > > - Shane, IPMC Member > https://www.apache.org/foundation/marks/resources >
I completely agree with Shane. To put it another way, podlings don’t make releases, the IPMC makes releases. In the past I’ve seen others advocate that one active mentor is enough for a podling. I do not share that opinion and think release voting is another reason a podling should aim (notice I’m not saying “must” here) to have at least three mentors. Mentors, by virtue of being more intimately involved with a podling, are much more likely to vote on a release than non-mentor IPMC members. To address the low level of IPMC members voting on podling releases, we need to understand why it happens. Is it just a lack of individuals’ time? Is it that IPMC members expect a podling’s mentors to review and vote on a release? Historically I’ve tended toward the latter, but my thinking has changed over time. While podlings have an explicit set of mentors, the IPMC as a whole can be thought of as an informal set of mentors. Reviewing and voting on podling release candidates is one way for any IPMC member to help a podling whether a mentor or not. -Taylor --------------------------------------------------------------------- To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org For additional commands, e-mail: general-h...@incubator.apache.org
