Re: first proposal for a new maintainer policy

[Philip Hands <phil@hands.com>]

>  ``We are a bunch of 300 developers who maintain 2000 packages. Since
>  everyone knows best how to solve upcoming problems and since everyone
>  always agrees with others--or in case of disagreement we have a
>  constitution--there is no need to specify who is working on what part. 
>  Everyone can modify any package at any time.
> 
>  Though, we do have some `policy' manuals, these are just for fun. It's
>  up to the individual maintainer to either follow policy or not.''

I've been keeping an eye on this discussion, and find myself smiling at the 
slightly disingenuous way that the two sides have been characterising the 
opposing view.

When described by the opposition, the two views come out as:

  A:  Policy is just for fun, and can be ignored on a whim.

  B:  Policy is more important than life, and any infringement should be
      punishable by expulsion from the project, or preferably death.

Now clearly, these are not the views that are actually held by either side,
which seem to come out as something closer to:

  A:  Policy should be adhered to, except where (in the maintainers opinion)
      it would be more appropriate to something else (on technical grounds)

  B:  Policy should be correct and up to date, in which case there should be
      no reason to allow exceptions, because things that are justified
      exceptions should be included in policy.

Is that fair ?  < donning flame proof armor ;-) >

These are not nearly as far apart as you guys are making out, and could be 
combined to say something like:

  Policy should be adhered to.
  
  In cases where the policy conflicts with what they consider to be best for
  their package, they can chose to ignore policy, as long as they also
  attempt to have policy changed by discussing it on debian-policy.

  If this discussion results in a change in policy, well and good.

  If the discussion concludes that they were wrong, they must fix the bug that
  they have introduced into their package by ignoring policy.

  While the discussion is under way on debian-devel, there is little point
  submitting bug reports pointing out the policy violation, unless that
  violation results in behaviour that could damage a user's system if they
  installed the package.

  In any case, if a maintainer insists on uploading buggy packages, against
  the consensus of the Debian developers, various sanctions, up to and
  including expulsion from the project are always available.

Of course, if the policy included a:

  ``Policy may by ignored while the clause in question is under discussion''

clause, then the policy could also a have a

  ``Policy MUST be obeyed at all times''

clause, since the exclusion would be in the policy ;-)

Does that satisfy both sides ?

Cheers, Phil.


--
To UNSUBSCRIBE, email to debian-policy-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org