incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Leo Simons <>
Subject PMC composition, inactivity, policy (was: Re: ActiveMQ and ServiceMix reports)
Date Fri, 17 Mar 2006 16:39:02 GMT
Can people *please* get into the habit of changing subject lines to
match subject matter around here? Thanks!

On Thu, Mar 16, 2006 at 12:03:27AM -0800, Henri Yandell wrote:
Alan wrote:
> > Not providing commit karma seems to be a bit like forced retirement
> > because of inactivity.  Something that ASF frowns upon.

We do have this concept of "emeritus". In some projects emeritus status is
supposed to be about automatic after 6 months of inactivity. If someone is
MIA then a good reason not to provide commit access (not neccessarily the
abstract privilege, but the actual access with login and password) is
security -- we don't want people not managing their passwords and/or SSH

> > Let's do another scenario.  Someone works very long and hard on one
> > component of the project.  That component becomes very mature and rock
> > solid so, we really don't hear from him very often.  Is it fair that he
> > doesn't get commit karma when it graduates?  IMO, no, it is not fair.
> > Is it fair that he does not make it into the project PMC?  Yes, it is
> > fair, IMO.

What if there is suddenly a problem with this component (perhaps a patent
claim) and the PMC has to act? What if he is the only person on the PMC
who really knows about prior art for this component? Would it be "fair" to
have him be inactive yet on the PMC mailing list for say, 3 years, prior
to that happening, if it means that the PMC as a whole is able to deal with
the patent claim within a week instead of within two months?

What is not fair is someone making decisions while not doing any of the
work (eg we think "meritocracy" is "fair"), but you can be on a PMC yet
not make any of the decisions (you just don't vote), and this is not really
a very bad thing, as long as you live up to your PMC responsibilities when
you have to.

Its not black and white. Every PMC is unique, as is ever project, and every
individual. In the end we're talking about individuals working together and
they do that in many different ways.

> I'm convinced - this definitely seems like a very good reason to have
> inactive committers following an incubated project through to either
> TLP stage, or into another TLP, but not being on the PMC. I'd be less
> convinced on a project that wasn't previously open so that there was
> no way to know if committers had previously contributed; but for open
> source projects who join the incubator this is a  great point.

*shrug*. This neatly ties into the whole "open source" vs "open development"
discussions. Some projects are more "open" than others.

> In fact you've helped me resolve a direction on my general problem in
> this area at Jakarta where we have a huge number of inactive
> committers (300) and PMC members (40) - inactive committers good,
> inactive PMC members bad.


As an example of inactive PMC members, consider excalibur. The excalibur
PMC hasn't had to do much over the last few months except help the chair
submit a board report or two and vote on a few releases. The entire PMC
(as a PMC) is inactive and this is not a bad thing. If there is something
needing PMC attention, all of a sudden enough of those people would jump
right back into active gear to deal with it.

I don't really understand where all the talking about what the right and
wrong policies are is coming from. The right policy is a PMC deciding by
consensus (or some other voting scheme, sometimes) what its composition
should be. It seems we're looking to policy as a way to "scale" our
processes, and I think that won't work.


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

View raw message