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

Re: Developers vs Uploaders

* Anthony Towns <aj@azure.humbug.org.au> [2007:03:15 04:22 +1000]: 
> If people don't do a good job as a "maintainer" they should have their
> priveleges removed fairly promptly; and if a developers recommends
> people to be listed as maintainers who turn out to be a problem, or if a
> developer just doesn't stay around to help them out when the need arises,
> they should stop being allowed to recommend people.

Who gets to decide this and what kind of process would be implemented?
What exactly would be a good job?

> In the long term, I think it would be sensible to have keyring-maint
> be a group looking after both this keyring and the developer keyring;
> and also to have the members of that team not be part of DSA or DAM to
> avoid concentration of powah.

While I get the idea of decentralizing power, to some extent, here are
my concerns with this proposal:

- It's not obvious what problems it's meant to be solving
  Is it meant to be a stepping stone for NM? Prevent sponsor(ee) burnout
  and boredom? Is it meant to replace NM eventually? If so, what are the
  current NM problems _besides_ sponsoring and boredom that it's
  solving? And what's causing all the waiting -- is it reasonable or
  not? Are the NMs at fault? etc.

  Because here's what I think about those: NM could be revamped, but I
  think the last thing it needs is MORE steps or more avenues with even
  more gatekeepers at each one. People will be in NM for five years if
  that happens. Which leads me to my next point...

- Overly bureaucratic

  NM is already too complicated. This is basically a way to implement
  middle managers in Debian, as far as I can tell -- it's one thing to
  have multiple points of failure, but if they're all sequential then
  there's no real gain, and finding ways to get a group of people to
  work together and agree (like with a proposed "court" of voters --
  eek) is not something I am personally convinced will work. The only
  time it works is with cabals, and they only work in secret, so I'm
  disinclined to believe this will function differently. 

- More power structures
  OK, this may be a personal bias, but I have a real issue with power
  structures and think as few of them as possible should exist,
  especially in Debian. IMO the only way to mitigate against the
  inevitable carving out of fiefdoms is to not implement such things. 

- Stratification

  As a subset of the power structure thing, one of the other issues I
  foresee is a "some developers are more equal than others" thing
  happening. I'm having a hard time thinking of how to explain this,
  because it's a bit "télétubby", as Joss would say, but I think there
  is ample world history that supports the idea that caste systems
  negatively impact societies, unless you're in the upper caste. And
  this bullet would not be complete with a gratuitous mention of
  Gattaca and A Brave New World. :) 

- Trust and upload rights

  I don't think upload rights should be given out trivially, but I also
  think that if you've got upload rights, you might as well have full
  rights and that you ought to have been through all of the "inspection"
  a current NM/DD would have to go through. The idea I'm getting from
  this is that you don't even have to agree with Debian philosophy or
  have much verification for who you are in order to have upload rights.
  Because if you don't, then upload rights probably shouldn't be given
  to you, and if you've already proven that stuff, then, again, you may
  as well be a full-fledged developer. And what happens when the DMs
  realize they can't vote (but want to) and that they now have to
  complete NM anyway? 

Overall, my impression is that this is a bandaid that might have the
unfortunate side effect of ripping our collective skin off. SCC is

off the chain like a rebellious guanine nucleotide

Reply to: