Re: [9.4 CF 1] The Commitfest Slacker List

On 7/3/13 7:25 PM, Bruce Momjian wrote:
> The extrapolation of Josh's approach is that committers
> have to do work that the community wants to maintain their commit
> rights, but their commit rights are helping the community, so why would
> people care if you take them away --- you only hurt the community
> further by doing so.

The main problem with having inactive committers (which I don't intend
to include the important subject matter committers, who I'll get into at
the end here) is that they skew the public information about who commits
in a counterproductive way. People visit
https://wiki.postgresql.org/wiki/Committers , sees that list of names,
and then make conclusions based on its content. And some of those
conclusions are wrong because the data is inconsistent. The standards
listed are applied when new committers are promoted, but they are not
applied symmetrically to remove ones who don't anymore.

The #1 obstacle to my getting more time to work on core PostgreSQL is
that companies presume regular submitters who are also non-committers
don't do a very good job. If you are competent and have a consistent
track record of contributions to an open source project, the project
would make you a committer, right? Conversely, if you've been
contributing for a while but aren't a committer, the most likely
explanation is that your work quality is poor. That is a completely
reasonable viewpoint based on how most open source projects work. The
really terrible part is that it means the longer you've been submitting
patches, the *less* competent you're assumed to be. When I tell people
I've been submitting things since 2007 but am not a committer, the only
logical explanation is that my submissions must suck very hard, right?

From that perspective, people who are listed as committers but don't
actively do work for the project are causing me a serious problem. When
someone who rarely commits can obviously qualify, that *proves* the bar
for PostgreSQL committers is actually very low to casual observers.
That's the message the project is inadvertently sending by leaving
committers on there if they stop working actively.

The main thing I'd like to see is having the committer list, and its
associated guidelines, updated to reflect that there are subject matter
experts committing too. That would pull them out of any "what have you
done for me lately?" computations, and possibly open up a way to get
more of them usefully. Here are the first two obvious labels like that:

Michael Meskes (meskes) - embedded SQL
Teodor Sigaev (teodor) - inverted indexes

When even Josh Berkus doesn't even know all of this information, it's
clearly way too obscure to expect the rest of the world to figure it out.

It also boggles my mind that there isn't already an entry like this on
there too:

Thom Browne - documentation

Each time Thom passes through yet another correction patch that is
committed with no change, I find it downright bizarre that a community
with such limited committer resources wastes their time with that
gatekeeping. The standards for nominating committers seem based on
whether they can commit just about anything. I think it's more
important to consider whether people are trusted to keep commits within
their known area(s) of expertise.

--
Greg Smith 2ndQuadrant US greg(at)2ndQuadrant(dot)com Baltimore, MD
PostgreSQL Training, Services, and 24x7 Support www.2ndQuadrant.com