Training and mentorship (Re: orphaning my packages)
On Wed, Nov 20, 2002 at 02:57:26AM +1000, Anthony Towns wrote:
> On Tue, Nov 19, 2002 at 11:01:40AM -0500, Matt Zimmerman wrote:
> > An excellent point, and one worth stating in big, bold letters on
> > nm.debian.org. NM is not meant to be a training program, but a
> > certification process of sorts.
>
> Would it make sense to have a training/development program?
>
> For new maintainers and experienced maintainers as well as / instead
> of applicants?
I think that the sponsorship system fulfills (or should fulfill) this need.
At least, that is how I interpret my role as a sponsor: not just as a filter
to screen out broken packages, but to act as a mentor. For applicants, this
kind of one-on-one volunteer mentorship bridges the gap between our training
materials (documentation such as the policy manual and developer's
reference) and the real world of packaging.
One way to extend the training paradigm to new and experienced maintainers
would be to incorporate a peer review process. It's not difficult to
envision a semi-automatic system for this, which would try to guarantee that
N developers had at least:
- read the changelog
- read the diff from the previous version
- installed or upgraded the package
before it is considered to have passed the review process. If
infrastructure were in place for this, it could be utilized by the
sponsorship system as well. We get into some sticky sociopolitical
situations, though, when we need to decide which packages require how much
review, and by whom.
> (There are ways to do such things other than telling experienced people
> "stop doing useful work, help these stupid people instead"; having less
> experienced people assigned to help more experienced people can pass on
> some knowledge and divide harder jobs, better than just leaving
> inexperienced people to work on some new package no one cares that much
> about; moving experienced people from one project they've been working on
> for a while to a new project that they're less experienced with can also
> have benefits -- letting new people gain the same experience the oldbie
> had, and letting the oldbie broaden his/her horizons. This is easier said
> than arranged, though)
Agreed, especially on your last point. In an ideal world, a work ethic and
a desire for personal growth would motivate all developers to refocus their
efforts from time to time, but in reality I'm sure that most of us suffer
from some degree of stagnation.
Stagnation is often best addressed by some external stimulus, but I can
think of no realistic strategy for connecting developers with new tasks that
has chance of actually working without requiring personal motivation on the
part of the developer.
--
- mdz
Reply to: