Board/Project Governance

Michael Scherer misc at zarb.org
Fri Sep 13 14:03:55 UTC 2013


Le jeudi 12 septembre 2013 à 16:44 +0100, Matthew Garrett a écrit :
> Fedora is fundamentally a somewhat anarchic collective. Our governance 
> model is built around fostering coordination rather than leadership. In 
> that context, the existing board structure is an obvious failure. The 
> meaningful things we do are make decisions on the use of project IP and 
> come up with handwavy ideas about project direction. We're theoretically 
> the last line of appeal against any other decisions, but in most cases 
> we're less qualified to make that decision than whichever group did so 
> originally.
> 
> Handling project IP is important. I don't think we need a board to do 
> it. Determining project direction is important. I think an 
> elected-but-not-actually-representative board is the wrong body to do 
> that. Being entirely honest, the reason I stood for the board was that I 
> was concerned about the potential for a refocusing of the project's 
> direction in a way that didn't interest me. It's great that people chose 
> to elect me, but right now I have no other involvement in Fedora's 
> governance. My contributions to Fedora are fairly minimal - I work on a 
> few tiny corners of some important packages, but otherwise I just talk a 
> lot. The fact that it was so easy for me to get on the board and 
> influence these decisions is arguably a failure in process.
> 
> So, a council-type arrangement seems entirely appropriate. I endorse 
> this proposal.

In the past, I worked on the mageia community to setup the governance,
and after lots of discussion, and based on our experience ( read
failures ) on mandriva, we did end with something like the proposal of
Josh.

Basically, 2 governing bodies were created. The first one is a board,
and it mirrored the kinda required structure needed for legal reason
( ie, legally responsible ). When I served there, the board was doing no
meeting or anything. The duty were mostly legal paperwork, financial
report, etc. We didn't had much things to say on the trademark, since at
that time, there was no problem. People were elected by association
members, then elected a president, treasurer, and a secretary ( required
by bylaws ). Election was among the association members, and to be on
the association, you had to be active.


The 2nd body was the council, with weekly irc meeting, and people from a
set of fixed team that we decided based on past experience. each team
choose 2 persons, to reduce the bus factor. 

More info on https://wiki.mageia.org/en/Org


Now, based on my experience as serving on basically every level of the
graph on previous web page, I can say we had some issues that should be
taken in account :

- Mageia board as a group was doing almost nothing, except minor
irregular tasks, or taking care of the yearly meeting in FOSDEM. All
people on it were also part of the Council, due to bootstrapping and
design of the gouvernance, so the overlap was quite obvious. Each people
were active, do not get me wrong, but I felt this was a issue, and that
the structure could have been simplified since it was obviously too
complex for most people to grasp. But we had to keep it for legal
reasons. 
Everything else was handled by Council, since this was basically the
same people, and IMHO, would be for a foreseeable future. 

So if we choose to have a board and a council, I think we should make
sure there is a need for the board. Mageia modelled the governance based
on the current french system, ie having senat and assembly, the assembly
taking the more active duty, the senat being here more to check
conformity ( roughly ). We (Mageia) wanted to avoid a too sudden change
of the association ( and the ressources ), so the divide was here to
retain agility on one side, and conservatism on the other. That's not
really a problem that would happen at Fedora, so maybe there is no need
for 2 groups. Hence I support the replacement of the board by the
structure of a council. 

- the set of teams, while being fixed arbitrarily, not in stone and good
enough (IMHO), was slightly unbalanced. It was easy to find packagers,
not so easy to find people for the web team. So in the start, we had
people from some team, and no one from the others. We used to have a
tool to make election in some kind of self service way, but it was not
good enough to be used by most people, so the size of most team was
small enough to just decide informally.

This in turn ask the question of "who should participate to council". I
think we didn't really articulate that the set of team was more a
minimal requirement, rather than "you must be in this team or shut up".
Now, having too much people on irc make it inefficient, but that's
another issue.

We already know who must participate to distribution go/nogo, and I
think that's a good start. But since that group would handle more than
just the go/nogo, so it would make sense to have groups that do not
participate in the current meetings to be present. For example, there is
no one from Legal or I18N. Mageia do have someone from security team,
but they do not appear at all in Fedora, because that's handled by
Security Response Team from RH. What about CWG, etc ?

The legal part is quite important, mostly because that's partially what
the Fedora Board do, and so if we replace it fully, maybe having a
separate team/group would make sense to take that part of the work and
it also address the fact that Board do not have the required expertise
to handle some request, by pushing those requests somewhere else where
there is. 


- weekly meeting are good and bad. Good, because they give the cadence,
permit to make sure we see that team are here. bad, because they usually
do not produce much, end up being a replacement for mailling list, take
lots of time, are hard to organize ( especially for timezone issues ),
etc, etc. 

But that's a hard problem. Not having irc meeting mean that issues are
discussed in a unaudited way, or on mailing list, which then go back to
"which mailling list" ( and people complaining about there is too much
ml, and/or too much mails ).

Still, i would push for having less irc meeting, to be more inclusive of
folks outside US. 1 or 2h of meeting can be lot when you can only devote
6 to 8h to a project.

So, to wrap up :
- yes from me to move to the proposal
- board could be replaced IMHO, and I think a different name would help
to clearly show the change 
- tasks that we feel the board were not adequate would be moved to a
team where it would, creating it if needed


-- 
Michael Scherer



More information about the advisory-board mailing list