Need advocate

[Joey Hess]

Michael S. Fischer wrote:
> Isn't the purpose of NMUs to cover for, er, "busy" developers?  I see
> people like Joey Hess doing this all the time.

NMU's are regarded with mixed feelings by most debian developers. That's
not suprising, since they break one of the main invarients debian is
built on: one package, one maintainer. There's always room for
exceptions to the rule, of course, but this rule has so permiated Debian
that it should not be broken lightly. There's almost always an underlying
subtext behind a NMU, or at least behind one of my NMU's. It might
include any of:

  - This is one of the packages whose maintainers have granted me
    unlimited NMU power without warning. There are a few of these, for
    various reasons.
  - I might be a good friend of the maintainer of the package,
    and so I know a NMU won't get on his nerves much. Or he asked me to
    look after it while he's away.
  - Even though we haven't gone to the bother of listing me as the
    maintainer, I am a co-maintainer of the package.
  - There is a general consensus in Debian that the issue I am NMUing
    for is important enough to warrent a NMU. FHS issues, for instance.
  - I've decided that someone is MIA, or that I just don't care if I
    piss them off.

NMU's typically result in some reactions on the part of the person whose
package was NMU'd, too:

  - So you think that I'm a bad maintainer?
  - You don't understand the details of my life/vacation/maritial
    status/fill in blank, quit being so impatient.
  - You just created a bunch more work for me, since I have to go
    integrate your NMU with the new version I was just about to upload.
    No, really, I mean it, I have the already-compiled .deb's sitting
    right here.
  - Did you just break my package? You probably don't understand it, and
    only I have the appropriate blessed sacrificial knife..

> I already sent a patch to the package maintainer; he has yet to
> integrate it.

Hmm, so have I. I'm not going to NMU dpkg (for instance) today though.
Far too few of the points in the first list, and far too many of the
points in the second apply.

> What's the point of uploading patches that never get committed?

"never" is a long time.

> I want to help Debian be a better distribution, but if I have to keep
> jumping through hoops in order to become an "official" developer, I'm
> afraid it's not going to be worth my time.  Isn't being able to fix bugs
> enough?

Two responses here:

a. There are many ways to help make Debian better without being a
   developer, and oddly enough, folks who engage in them before applying
   often seem to slide through the process like a greased pig[1]. The
   ability to upload strings of bits to a server somewhere is pretty
   unimportant in the scheme of things, and most every other way of
   contributing to the project is open to everyone. For example, if you
   think the maintainer of whatever the package is is MIA and that he's
   been ignoring the clearly correct patch to fix whatever the bug is,
   well you're free to post to that effect to debian-devel just like any
   developer would.
b. People are going to react a lot better to "I want to join Debian to 
   help with Debian QA". Than they are to "I want to join Debian so I
   can make lots of NMU's". Even though the former just requires a
   smidgeon of a wider view than the latter, and they otherwise mean
   pretty much the same thing. No aplogies, that's just the way it is.

-- 
see shy jo

[1] There's a rule that I have to meet my quota of such folksy
    metaphores each day, now that I'm an official hillbilly hacker.
    Sorry if I'm not too good at it.