Universe Contributors application for James Westby (james_w)

[Michael Bienia]

On 2008-08-01 14:00:47 +0100, James Westby wrote:
> P.S. In the last paragraph I hinted at a slight problem I see
> with the current system. If someone does a lot of good work through
> the sponsors queue then they may not interact much with other
> developers, and so it may be difficult for them to know who
> to approach to advocate them. An application just with evidence
> of their work may be good enough, but it's not really how things
> seem to work for most people. It seems like something that
> the MOTU Council may want to consider.

I also noticed this problem; from both sides: as a sponsoree (for my
main uploads) and as a sponsor for universe uploads.

I do my 'main' uploads through the ubuntu-main-sponsors queue and I'd
need to collect a list first to see who sponsored me and how often and
would be a good candidate for a comment for a core-dev application.
I guess my uploads got sponsored through several core-devs and not just
a few, and most didn't got enough uploads to build an opinion.

While working on the ubuntu-universe-sponsors queue I usually don't look
on the name of the person getting sponsored. I just look at the contents
of the "bug" and decide if it's of good quality to get sponsored.
It happens then sometimes that I get contacted from a potential
applicant and get asked for a comment. I remember seeing that person in
the u-u-s queue but not which uploads and how many uploads I actually
sponsored from the queue. Sometimes it just a few (and also spread about
several weeks) which makes it very hard to build an opinion on the
technical/process skills without digging in the other sponsoring bugs
and re-evaluating the sponsoring requests again.

Currently I don't have a solution for this. Assigning someone a small
set of sponsors isn't a good idea as they might get busy or not
available for a period of time (work/private live/vacation/etc.) or the
packages the sponsoree is working on fall into different "classes" where
one needs different sponsors who knows the matching subpolicy to better
judge the correctness of the contribution, e.g. someone working first on
a KDE package, then a Gnome package, then a Perl package and finally on
a Python package. It seems impossible to determine such a set of
sponsors before the need for sponsoring arises. Changing it midway
doesn't help as one ends with many sponsors who didn't see enough
contributions to build an opinion.

Michael