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

Re: Re-Proposing: General Resolution on Init Systems and systemd

Sam Hartman <hartmans@debian.org> writes:

> I have high confidence that a decision in this space will give the
> policy editors the tools they need to break the consensus deadlock
> without having to resort to overriding them or setting policy as a
> project.  That is, in this case, a non-binding statement that we know
> the project supports will be sufficient to unblock things.  So I've
> proposed several such non-binding statements that the project could
> make.

I will go a bit farther than that as a Policy editor.

Consensus isn't an inherent requirement of Policy.  It's a decision we've
made about how to make changes to Policy for several reasons:

* Policy is the most useful for the project when it has the general
  support of the project behind it, and becomes less useful when it
  becomes controversial.

* We're very aware of the fact that only a small fraction of developers
  participate in the process, so one objection is reasonably likely to
  reflect the opinions of a lot of people who aren't involved.

* Most things in Policy are not urgent and we can take the time to get
  them right (although I would say this has currently gone too far towards
  being slow for mostly other reasons).

* Resources for Policy work are quite limited, and arguments and hostility
  drain those resources dramatically, so it seems wiser to primarily use
  our resources on things that do not provoke large arguments.

If the project decides a matter of policy by GR, that is a far more
effective way of addressing all of those concerns than the consensus
process.  We know as clearly as it is possible to know what the general
position of the project is, everyone has a chance to participate and the
discussion is much more prominent than any Policy discussion normally is,
and the GR process generally does a good job of exploring the relevant

Therefore, if the project decides something by GR even as a nonbinding
statement, it would take a *lot* of convincing for me to not change Policy
accordingly.  Basically someone would have to convince me that there is
some fundamental technical impossibility or some significant set of facts
that the project didn't have available when considering the GR.

> I also don't think it is appropriate to consider something overriding a
> delegate unless it is overiding a specific decision of a delegate.

For the record, it's not possible in this case to override a decision of
the delegated Policy Editors on this topic, since we haven't made one.

It would be possible to make a decision that is within our delegation
scope by GR.  As with NMUs, I'm very happy for the project as a whole to
fix my bugs for me any time y'all would like.  :)

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

Reply to: