incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Roman Shaposhnik <>
Subject Reflections from the outgoing Chair
Date Thu, 01 Jan 2015 01:26:37 GMT

when a honor of the IPMC chair was bestowed on
me beginning of 2014 it was crucial  that the
position remains to be rotated among IPMC members.
As an aside: while Marvin felt that the ideal rotation period
was 6 month my personal belief was that 12 months
makes it long enough to be able to accomplish something,
while short enough to still accomplish the goals of a
rotating chair.

At any rate, with my 12 months almost up, it makes the last
day of the year a perfect  cut-off point to start talking about
transition the Chair position, if it wasn't for one issue:

At this point I'm really convinced that what ASF needs is
not the next IPMC Chair, but the *last* IPMC Chair. That
is to say, I truly feel like the best outcome for everybody
involved in Incubation process is if by the end of 2015 IPMC
gets completely dissolved. Here's why:

First of all, as was pointed out in two other threads by Chris
and Benson (see the quotes bellow) the current process lacks
the most crucial bit of what Incubation is supposed to be all
about: Apache project on training wheels.  Instead of teaching
our podlings what it really feels to have a responsible PMC
and a Chair skilled in the "Apache Way" dealing with the board,
we have IPMC.

After serving in my current position for almost a year, I'm fully
convinced by now, that there's a very fundamental problem
with the organization. The existence of IPMC (despite all
the goodness that still comes from it) has become a too-convenient
of a excuse for *everybody* to play a really nasty shell game with

While the situation with ASF TLPs varies, at least the system
is setup in such a way that there's a very clear accountability
What's more important, there's a vested interest from those with
authority (PMC, PMC Chair and the Board) to make sure the
project is doing the right thing. Presumably, if you're on PMC
of a TLP you really don't want the PMC to be dissolved and
the project go away. You're not a member of some ethereal
"league of extraordinary gentlemen" (aka IPMC) your status
is in direct relationship to the livelihood of your project. Without
the project there's no status, which is exactly *not* the case
with IPMC. A pretty powerful motivator for doing the right thing
is clearly lacking.

Now, one might say that we don't need to dissolve the IPMC
in order to fix this, one might say that something along the
lines of original Ross' proposal would do. I would disagree. I think
that the only way to send the message or clear responsibility
is to make it impossible to be associated with an incubating
project in any other way, but being on its PMC. You're either
in or out. There's no other place to boost your ASF karma
(which, sadly, I've seen around IPMC more than I'd like to).

But wait, there's more! This real assignment of responsibility
wouldn't just happen at the mentor level, it'll extend all the
way to the board. The board will be directly engaged in
overseeing the incubating projects and that direct interaction
will be as much a part of the Incubation experience as
producing releases or growing the community. The scalability
of the board is not an issue here. First of all, the board still
needs to read all the Incubating reports and moreover
if the board doesn't feel scalable enough today, why would
all of a sudden scale if all the projects vote on graduation
at once tomorrow (and pass!)? If nothing else, this real
engagement gives the board a very early indicator of
its own scalability issues if any. Early warning are a good
thing, not something that needs to be feared.

All in all, it feels like direct overseeing of Incubating projects
would be a good thing for the poddlings, a good thing for
the mentors, a good thing for the board and ultimately
the only mature and responsible way of making sure that
the project we try to embrace get the best shot of becoming
ASF TLPs. At this point, I'm struggling to see any potential
negative effects. In fact, if there's one thing I really would
like everybody commenting on this thread to focus on it would
be that: arguing for potential downsides.

With that, I'd like to thank all of my IPMC colleagues for
this great opportunity and wish all of you the Happiest
New Year!


From: Mattmann, Chris A

It’s not just the board - again please see the table I’ve listed
at the bottom of the wiki. What my proposal does is remove the thinly
veiled “IPMC” as the “catch all” which in fact doesn’t catch all. On
its 150+ person committee - I supposed there are < 20 active people
who keep showing up. I have statistics to prove it (see my active
mentors tool I’ve shown) - I have experience having mentored many
podlings to prove it; and the mailing threads prove it. So, promote
those 20 people to ComDev PMC, promote them to ASF members, promote
them however, my guess is that they *care* about the foundation; we
want these people helping new projects, and they will continue to
help those new projects - along with the board - along with everyone
From: Benson Margulies

Here is where the 'Mentors in the Project' (whether directly reporting
to the board or not) leaps up and looks like a great idea to me. The
whole goal of incubation is to run an Apache project on training
wheels. How does an Apache project run? WIth a chair and PMC members
supervising it and _reporting to the board_.  The proposal, as I see
it, is to tell the champion and other mentors that they, and not the
entire IPMC in some nebulous fashion, are the PMC in the PPMC. By the
time the podling graduates, their need to have expanded themselves to
a larger group.

The board may choose to keep the IPMC around to organize and support
this process. The board may choose to continue to ask the IPMC to add
an extra layer of supervision. But the heart of the proposal is to
insist that every podling be nucleated around at least three people
who have the experience to operate as a PMC and have volunteered for
the responsibility.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message