Server WG Governance

[Stephen Gallagher]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 04/21/2014 03:42 PM, inode0 wrote:
> On Mon, Apr 21, 2014 at 2:08 PM, Stephen Gallagher
> <sgallagh at redhat.com> wrote:
>> At the risk of starting two out-of-control threads in one day, I 
>> wanted to bring up a discussion I had with Karsten Wade last week
>> at Red Hat Summit. We were discussing various governance
>> methodologies, specifically that of FESCo, the Board, the WGs and
>> CentOS.
>> 
>> There are some very interesting ideas that the CentOS Board has
>> put into place, the most relevant I think is their mechanism for 
>> consensus-based decision-making[1]. I'd like to describe it a
>> little bit here and note how I think in many ways this is pretty
>> much how we in the Server WG have actually been operating thus
>> far and that we may want to actually formalize it.
> 
> I think some variation of consensus decision making would be a
> very nice fit. While I am a pretty big fan of the process in
> general I'm normally not so fond of using a unaminous voting rule
> but that is a detail you can think about. Other groups do use other
> less absolute voting rules, like unanimous - 1, which allows for a
> small bit of dissent without derailing the decision.
> 

The problem with that approach (as I see it) is that it doesn't
address the marginalization problem. The point of a consensus vote is
that it ensures that no member of the group is being ignored (or if
they are, it's ultimately because they themselves have refused to meet
in the middle). Consensus - 1 still leads to "sides" and just moves
the target.


> The one abstract concern I have about consensus decision making is 
> that it might tend to water down the difficult decisions that
> really need to be made at times leaving the governed body with
> either the status quo or something quite different from what was
> proposed to begin with.
> 

I'm actually not sure that's a bad thing. If a "hard decision" needs
to be made, and the group can't come to consensus about it, that says
to me that the decision wasn't fully thought-out. There is always a
slight risk that eventually we'll miss out on a major shift, but I'd
like to think that the process is sufficiently lightweight that we
would be able to return to the question and answer it differently.

And if the result is to get something quite different from the initial
proposal, that says to me two things: 1) The initial proposal wasn't
good enough and 2) nine intelligent individuals ultimately decided on
something they could all live with. Knowing what I know about Fedora
Contributors (namely that we are all ultimately on the same side and
want to see Fedora and FOSS succeed and thrive), I can't really see
any alternative decision going too far afield of what we hope to achieve.


> I'll be watching with great interest as I too have considered 
> proposing something based on consensus decision making for one of
> the other governance bodies in Fedora.
> 

This is the tricky part. A consensus-based model is *very* difficult
to shoehorn into the system after the group has been established for a
long time. I think we have an opportunity to do so in the Server WG
right now because at least so far, we've managed to reach consensus on
almost everything anyway; we haven't really had to deal with splitting
on major decisions and fighting for swing votes.

But with groups like FESCo, the Board, FPC, etc. (not singling any
out), there's a long and entrenched history of majority-based
decision-making. The unfortunate side-effect there is that it has on
some occasions led to individuals or factions being unable to work
together (which is ultimately disfunctional).

I think that if such a system was to be implemented, it really needs
to come close to the beginning, before such enmities have formed (or
else the effect will likely be that no decisions are ever reached,
because neither consensus nor the emergency ban will be possible).

Certainly, if any of those groups want to move towards it anyway, I'd
happily support them on it.



As an aside, Karsten has put up a blog post (inspired by this
discussion) about the consensus model. I'll shamelessly plug it for
him here:
http://iquaid.org/2014/04/21/why-consensus-decision-making-is-better-for-open-source-projects/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iEYEARECAAYFAlNVeB4ACgkQeiVVYja6o6MCcgCeM0k46D77+MHjK/RAJhShG6TN
nacAoLCnTNkAmGMstymsPB+q7kXjrpPo
=giMc
-----END PGP SIGNATURE-----