incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Stein <>
Subject Re: my pTLP view
Date Mon, 26 Jan 2015 08:46:24 GMT
On Sun, Jan 25, 2015 at 4:01 PM, Dave Fisher <> wrote:

> On Jan 25, 2015, at 1:22 PM, Benson Margulies wrote:
> > On Sun, Jan 25, 2015 at 3:53 PM, Bertrand Delacretaz
> > <> wrote:
> >> On Fri, Jan 23, 2015 at 8:58 PM, Greg Stein <> wrote:
> >>> ...They are reporting to the Board. We know what inactivity looks
> like. So we
> >>> ask the PMC to fix it, or we shut them down....
> >>
> >> I know how that works, it's just that with your pTLP proposal the
> >> podling is "at the mercy" of their mentors - if the Incubator PMC
> >> disappears it might be hard for them to recruit initial or new
> >> mentors. Not a blocker for me, just an observation.
> >
> > I want to be clear about the hypothetical here. I think it is, "The
> > board establishes a PMC containing some people whom it knows and
> > trusts, and there is a larger community of some other people whom it
> > would like to get to know and trust. (Or, even, the board included
> > some of the second group in the PMC at the outset.) Before the second
> > group merges with the first group, the first group loses motivation
> > and disappears."
> >
> > It seems to me that this is not likely. To me, at least, signing up to
> > be a PMC member is a much clearer commitment than signing up as a
> > mentor, and, while I might be distracted for a month here and there,
> > I'm not going to just wander away. I think I'd be pretty much typical
> > (in this one tiny respect) of anyone that the board for a new PMC. To
> > get into the pickle Bertrand is musing about, more than one of the
> > group has to wander off, so that the remainder are not available to
> > help recruit some successors from the general membership.

Very well-stated, and very clear, Benson. Thanks. Yes... that matches my
own thoughts precisely. The "failure mode" based around absent PMC members
is very unlikely.

But carry it a bit further: one of those is the VP. An Officer of the
Foundation. If that person disappears, it is quickly obvious, so we don't
need to worry about this case (easily noted and fixed). If the other
members of the PMC disappear ... well, that VP can *still* appoint people
onto the PMC, to get the roster back up to (3) active people who can +1
releases, and can acknowledge that the commits/direction of the project
comport with the PMC and the Foundation.

[ for those not aware: yes, the VP can unilaterally add people to the PMC;
no need to have 3 actives; so a reboot is always possible if a VP is around

> +1.
> I think that in this pTLP proposal there are no IPMC Mentors. These are
> not needed. Why? The Apache Members are coming in as the PMC. This is a
> much more serious commitment than being a Mentor. The pTLP is not an IPMC
> entity.

In the model that I have proposed: correct. This is simply another TLP
which the Board has mandated the "probationary/provisional" label upon.
Even the "requirement" for only ASF Members is a suggestion. A community
can arrive with any initial list of PMC members, but speaking with my
Director hat: no, the initial list should be small, greybearded, and
well-known. They can make an argument for one or two others, I'd think, so
the "rule" is more like advice on how to get the Board to approve it :-)

While we don't like BDFL's or tech leads at the ASF, many incoming projects
have individuals that fit into such a role. I'd expect that person to get
onto the initial PMC list. But I would never approve without (3) or more
ASF Members on the list.

> Incubator life cycle for a pTLP.
> - Proposal to be a pTLP.
> - IPMC recommendation to the Board as a possible fit to be a pTLP.

Sure. Or a community can directly approach the Board. However, the Board is
really bad about engaging a community in discussion, so the approach kind
of needs to be packaged with a high chance of approval (since back/forth
won't happen). That generally means an individual Director would work with
the community to put the proposal together (based on their impression of
what the Board would accept, and their own part in discussions during the
Board meeting to make it happen).

> Board life cycle for a pTLP.
> - Board accepts - they can accept no matter what the IPMC opines.
> - Board manages it like a TLP. Appoints Board Shepherd, etc. Not the
> IPMC's responsibility any longer.
> - Board decides when the probationary period is over and the little "p" is
> removed.



  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message