decision making [was Re: default theme consensus]

[Mark McLoughlin <markmc redhat com>]

Hi,

On Thu, 2004-08-19 at 02:56, Ryan McDougall wrote:

> And for the last time, I will disagree, and state that consensus is
> good/necessary but the buck needs to stop with someone when the thread
> dies. There is no such thing as 100% agreement, but where do 10+ people
> draw the line of "good enough" on the same issue?

	This is not a new problem, and we've tried various ways of "fixing" it.
One way we tried was the GEP[1] but that didn't work out because
everyone felt it was too process driven and took the fun out it. Also,
it didn't actually solve the problem because most of the GEPs never
reached a conclusion and are still in the "pending" state.

	You could argue that the release-team was another attempt to solve this
problem (at least for some decisions). By picking a bunch of people who
takes responsibility for making the tough calls about the release
process that no-one wants to make, you ensure that indecision doesn't
kill our release process. Personally, I think the release-team has
caused us all to get worse at making decisions because we assume that
either the release-team should be making the decision or there should be
an "authority" to make the decision.

	The reality is that the GNOME community will never stand for a
"benevolent dictator" or a body which makes technical decisions. We
prefer open discussions, inclusiveness and "rough consensus and working
code". We want it so that if you contribute to GNOME, your say counts.

	I think, at its core, the GEP process was just an attempt to formalise
what we already do. A person with some legitimacy in that area takes the
responsibility for getting the decision made. That person picks a small
but varied group of people whose opinion on this he feels will be really
valuable. He then opens it up to the masses for discussion and tries to
steer things to a consensus. When he feels there's a rough consensus -
e.g. the small group of people and most everyone else agrees too - he
makes it known that he feels the decision is made and everybody moves
forward on that basis.

	This how we just made the new modules decision. This is how a
freedesktop.org spec gets written. It can work, but the problem with the
GEP process was that we tried to formalise something which people prefer
when its completely informal.

	I guess all I'm saying is that people who feel strongly that a decision
should be made about something should take responsibility for driving
the process from beginning to end.

Cheers,
Mark.

[1] - http://developer.gnome.org/gep/gep-0.html