Re: [gentoo-dev] Recruitment process is moving back to quizzes

[Rich Freeman <rich0@gentoo.org>]

On Sun, Jul 15, 2012 at 10:32 AM, Peter Stuge <peter@stuge.se> wrote:
>  less idiotic
> idiots such as perhaps myself need years because we're doing
> whatever work as opposed to learning foundation bylaws by heart.

Well, I don't think the bylaws are a terribly important topic for the
quizzes, and unless something has changed I don't think they were a
topic in the past.  Sure, developers should understand the role of the
council and the trustees, but that doesn't mean that they need to be
qualified to BE a trustee.

However, I think that it is important to include a fair bit of "meta"
in the quizzes.  In fact, I'd consider this almost more important than
the technical content.  A developer who doesn't understand some nuance
of ebuild development, and recognizes this, and therefore acts with
maturity in asking for help and review before doing commits isn't much
of a danger to the distro.  A developer who is a technical wizard and
creates bots that do massive tree-wide commits to correct some
perceived problem without gaining consensus from the community is a
danger to the distro, even if most of the time they are completely
right.  I think it is more important that a developer be able to work
with others and recognize their own limitations, than to worry about
what those limitations are exactly.

When I look at most of the issues impacting Gentoo over the years,
rarely are they caused by some bug in an ebuild.  They happen, and
they get fixed, and usually the impact is very minor.

What really causes havoc around here is when people change ebuilds
without consulting with the maintainer, or when they go tweaking
system packages without a great deal of care and being part of the
appropriate team, and so on.  Lack of respect on mailing lists has
caused no small number of problems either.  Many of these issues have
dwindled in recent years, and I think it is precisely because teams
like the recruiters have been paying more careful attention to them.

Anybody can write good code.  You don't need to be a Gentoo developer
to do that, and if somebody lacks maturity and social skills they're
probably better off doing their work on the side with a proxy
maintainer pulling it in.  Calchan had both, and he still ended up
being more successful with OpenRC from the outside.  The KDE team has
in the past made use of bleeding-edge portage (or even non-portage)
features in overlays for development purposes, driving PM development
in the distro to yield an improved result when everything got merged
back in.  The bottom line is that you don't need commit access to do
great things with and for Gentoo.

What the Gentoo devs really need to be about is making all that great
code work nicely together.  That requires coordination and an eye to
quality.  That requires working well with those who differ in opinion.

That said, I don't want to diminish the importance of technical skills
too much.  I think Gentoo has created some really good infrastructure
that rivals what has been done with much larger distros.

Rich