Hello, VXQuery community,
Everyone here is well aware of how long it took to get enough "binding" IPMC
votes to approve your release. Unfortunately, this is not the first time
that an Incubator podling has experienced such difficulties.
The twin problems of binding vote scarcity and Mentor attrition have plagued
the Incubator for much of its history, and several of us on the Incubator PMC
have been trying to improve things for years. Finally, we may be on the verge
of a breakthrough: consensus support has been painstakingly assembled for
running an experiment where a podling's contributors are allowed to earn
binding votes.
I believe it would be to the benefit of all if VXQuery were to serve as the
test case.
Here's the current proposal, a variant on the "pTLP" option discussed earlier
but run entirely within the scope of the Incubator:
http://s.apache.org/atG
1. The initial PPMC shall be composed exclusively of IPMC members.
2. PPMC votes are binding for every release except the first.
3. One IPMC vote is required for each release after the first.
The conditions assume a podling which is just starting out. Here's the
proposal, adapted for VXQuery specifically:
1. The VXQuery PPMC shall be reconstituted to consist solely of Incubator
PMC Members Ant Elder, Marvin Humphrey, Till Westmann and Jochen
Wiedmann.
2. Additional VXQuery PPMC members will be voted in using existing
Incubator procedures.
3. Votes cast by VXQuery PPMC members shall henceforth be considered
"binding" for VXQuery incubating releases.
4. One +1 vote must be cast by a member of the Incubator PMC in order for
a VXQuery release vote to pass.
Of course we would immediately set about discussing nominees drawn from the
list of PPMC members as it exists today, and it is highly likely that we will
end up with the same group of people. However, we will also be discussing
what each nominee brings to VXQuery in terms of Apache governance expertise.
I would expect to participate actively in these conversations (though I hope
others will take initiative with nominations and such).
There are two benefits to VXQuery for participating in this experiment.
First, every PPMC member who gets voted in this way gets a binding vote on
incubating releases, so you will probably not have trouble getting them
approved ever again. But let's be straight: with Till now on the IPMC and
Ant, Jochen and myself now on board as active Mentors, you probably were not
going to have difficulty anyway -- so binding votes, as useful as they might
have been for the last release in theory, are not really that big a deal any
more.
The second benefit is that the exercise of rebuilding the PPMC with an eye
towards governing effectively as an Apache TLP will serve as intensive
education in the Apache Way. In my view, this is much more valuable, as it
will not only help us to assess the readiness of the podling for graduation,
but to fill in many gaps over the course of a running conversation. In the
July report, VXQuery asked for more active mentoring; this process would
deliver just that. When it concludes, I expect that a plan will have
coalesced leading inexorably towards graduation and beyond -- to a robust,
healthy life as a TLP.
But beyond the direct benefits to VXQuery, I hope that you will weigh the
considerable benefit to the wider Apache and Incubator communities of your
participation: if this model works and we can get it adopted as Incubator
policy, you will be helping to spare future podlings from some of the
frustrations you've experienced.
If you are amenable, I will present the proposal on general@incubator. Feel
free to cast votes. :)
But also please feel free to close out an "ordinary" incubation. It is my
opinion as Mentor that VXQuery's small PPMC needs to beef up its Apache-fu,
but that can be done either using the structured framework of this experiment
or in a less formal, unstructured environment. At Apache, we believe that
distributed self-governance is essential, so the choice is yours.
Marvin Humphrey
