Re: Four days
On Tue, Oct 05, 2010 at 12:14:36AM -0400, Michael Gilbert wrote:
> On Tue, 5 Oct 2010 13:52:09 +1100 Matthew Palmer wrote:
> > On Mon, Oct 04, 2010 at 10:32:24PM -0400, Michael Gilbert wrote:
> > > On Mon, 4 Oct 2010 17:37:19 -0700 PJ Weisberg wrote:
> > >
> > > > On Mon, Oct 4, 2010 at 1:15 PM, Michael Gilbert wrote:
> > > > > On Tue, 5 Oct 2010 06:30:22 +1100, Matthew Palmer wrote:
> > > > >> Yeah, that's a great idea! We should setup a mailing list where they can
> > > > >> get together and ask questions of each other and request someone to sponsor
> > > > >> their packages!
> > > > >
> > > > > What's so crazy about that? What would be so wrong with empowering
> > > > > mentees to help themselves?
> > > >
> > > > I think you missed his point. ;-) Such a list already exists. It's
> > > > called debian-mentors.
> > >
> > > OK, I see the attempt at irony now; although that really misses my
> > > original idea, which is to revamp the mentoring process with more of a
> > > team-based focus.
> > On the contrary; what is different about a team of people within Debian who
> > wish to assist and mentor potential new developers, as opposed to the
> > membership of the debian-mentors mailing list?
> A lot. The current process is individualized mentorship, not team
Not to my knowledge. Whilst some new maintainers may have an invitation
from certain DDs to e-mail them privately with their questions, in general
the intended process is that questions are asked on this mailing list, and
answers are given and discussed by anyone who feels qualified to answer --
DD, DM, "mentee", or interested bystander.
> One is to use the mentors mailing list as the maintainer for mentee
> packages. That way the burden of quickly orphaned packages is dispersed
> over the whole set of mentors rather than just one. Perhaps that will
> encourage more DD participation since they won't stick themselves with a
> lot of orphaned packages.
To clarify: the intended point of this proposal is to solve the perceived
problem that DDs don't sponsor packages because they're concerned that
they'll end up taking responsibility for a package if the maintainer ups and
I don't actually see that as a problem. There are simple ways to deal with
orphaned packages, regardless of the way the upload was made, and they work.
If a package I sponsor is abandoned by the maintainer, it gets NMUed,
orphaned and assigned to debian-qa like any other, and is then available for
The variant of this problem I do see, however, is the uploading of
surely-soon-to-be-unmaintained low-quality or near-duplicate packages,
clogging up the archive and making extra work for debian-qa et al. *That*
problem isn't going to be solved by changing the maintainer, it's only going
to be solved by not uploading the surely-soon-to-be-unmaintained low-quality
or near-duplicate packages in the first place.
On a practical point, making d-mentors the maintainer would clog the list
with large quantities of (mostly) bug-related e-mail, a la debian-boot,
making the list far worse for discussion. However a separate mailing list
could be created to avoid that problem (at the cost of requiring people to
subscribe to the other list, splitting attention, etc).
> The other idea is to reduce DD involvement in the mentoring process
> itself by making mentees more responsible for themselves. Take a set of
> mentees, have them work together to get their packages in shape, then
> maybe once a month (or every couple weeks) have them show the set of
> packages that they have ready to the mentors list. That would also
> reduce RFS traffic on this list. This list would become more of a
> coordination point for joining mentee teams.
There's nothing stopping that from happening now on this list. I don't see
that "batching" RFSes is going to either (a) reduce RFS traffic (because
nothing's stopping people from still posting them here, and even the batch
will have to be sent out some time), or (b) improve sponsorship rates (in
fact it'd probably decrease them, because checking 1 package a day every day
is far less daunting than checking 14 packages once a fortnight).
However, if you want to give it a go, don't let me (or anyone else) stop
you. Take Asheesh's lead and just start something, don't ask or wait for
official endorsement of your idea, because it will never come. Do it, and
if it works it'll catch on, and if it doesn't then... try something else.