[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: On the "Debian Maintainers" GR

Raphael Hertzog <hertzog@debian.org> writes:
> On Thu, 26 Jul 2007, Russ Allbery wrote:

>> My primary worry about this proposal is that it leads to an increasing
>> number of packages in the archive that are someone's pet project,

> Every useful software is the pet project of someone. :-)

> DM don't decide of new packages. It's the sponsor and the ftpmaster team.
> It's up to us to restrict what we upload.

I don't want to restrict what we upload.  The problem with something being
a pet project isn't that someone cares specifically about that project or
that it's not that broadly applicable.  It's when the person cares about
that particular package *more than Debian as a whole*.  Part of becoming a
Debian Developer is that understanding of how Debian functions as a whole,
and while it can still happen with DDs, it's a lot less likely, IMO.

One of the consequences of not caring about Debian as a whole, just a
particular package, is that transitions and similar global changes tend to
get blocked and held up by those packages, require NMUs, and so forth.

>> maintained by someone who isn't keeping up with Debian-wide changes, and

> Why that ? I expect all DM to be subscribed to d-d-a and would suggest a
> check (or even some enforcement with auto-subscription if we really want).

What motivation do they have, though?  They're only maintaining these
particular packages, and they've indicated via process that they're
interested in that *rather than* becoming a DD.  Whatever we say about
what they're required to read, it's going to be a lot harder to convince
them to care about project-wide issues than DDs.

>> At least with sponsored packages, the sponsor in theory should be
>> paying attention.

> That theory doesn't match my experience.

I'll be the first to admit that there are problems with our existing NM
and sponsorship process.  I really don't like the way that this proposal
addresses them by creating a completely new process with different
problems, though.  It feels like giving up on the existing processes,
saying "they're just unfixable and we can't get anyone to work on them,"
throwing up our hands, and making something new.

As tempting as it is, reimplementations from scratch are almost never the
right solution to a problem.

Because of that, any argument for DM that reduces to "NM isn't functioning
properly" or "sponsorship doesn't have enough checks" isn't convincing to
me.  I'd need to see arguments that defend DM as a good idea even in a
world where NM is working as well as possible and we can tackle problems
with the sponsorship process.  Otherwise, it feels to me like we're just
dodging the root problem, and I'm suspicious of adding complexity to an
area where we already have issues.

So far, the only arguments I've seen of that type are "I don't want to be
associated with the project but I still want to maintain Debian packages"
and "I don't want to go through the NM process just to maintain a single
package."  I'm sympathetic somewhat to both of those arguments, but I
don't think they're compelling.

Russ Allbery (rra@debian.org)               <http://www.eyrie.org/~eagle/>

Reply to: