GNOME hits a decision point [LWN.net]

There is a vibrant debate underway within the GNOME project. Certainly, there is no shortage of issues that GNOME could be discussing, such as, say, how to shoehorn a panel applet into less than four or five megabytes of memory. The issue at hand is not quite so exciting, however; instead, the GNOMEs are voting on a referendum which would reduce the GNOME Foundation board of directors from eleven to seven seats.

One might think that the size of the board does not matter that much, but a number of developers seem to have strong feelings about it. The arguments in favor of the change include the following:

There are not currently enough people in the GNOME community who are both interested in serving on the board and qualified to do so. A number of people who have run for board seats have put their names in only because they feared that there would not be enough candidates, or that unqualified people would be elected by default.
The eleven-member board is too large and unwieldy to get things done. Decisions get pushed back indefinitely, and nobody takes responsibility for making things happen.
Some members of the current board are not actually doing anything beyond warming their seats.

On the other hand, those opposing a change in the size of the board argue that:

The GNOME community is large, vibrant, and diverse. With such a talented group of people, there cannot possibly be a shortage of developers qualified to sit on the board. Rather than shrink the board, the GNOME community should actively encourage more of its members to seek election.
If the board is sometimes contentious, that's because the community as a whole is global and diverse. Obtaining a consensus on controversial issues will never be easy. Any consensus reached by way of shrinking the board (and thus excluding minority members) will be false at best.
If there is a problem with the board, it's in how it operates, rather than its size. Any effort to fix the board should look at other solutions, such as assigning specific responsibilities to each member. Jeff Waugh has also suggested a more executive-style governance structure with specific elected positions.

Voting is happening now, with the preliminary results scheduled to be announced on November 5. The highly scientific LWN "read the mailing lists" poll has concluded that the "no" votes are currently leading 55% to 45%, with a margin of error of about 30% or so.

In many ways, free software projects continually break new ground on governance issues. A quick survey of high-profile projects shows a wide variety of governance structures, including dictators (usually but not always benevolent), oligarchies, corporate fiefdoms, extreme direct democracies, and more. GNOME has a sort of representative bureaucracy which is seen to be in need of some tweaking. The GNOME hackers will certainly continue to produce code regardless of the outcome of this particular referendum. But the overall effectiveness of the Foundation can have a long-term effect on how the project is perceived and which technical directions are taken. So it will be interesting to see how this project ultimately decides to govern itself.

(Log in to post comments)

FEWER = BETTER

Posted Oct 27, 2005 4:58 UTC (Thu) by atai (subscriber, #10977) [Link]

Not necessary true.

A smaller number means the decision will be in the control of fewer people. It is more likely to make decision that alienates many people. A large board decrease such possibility.

FEWER = BETTER

Posted Nov 3, 2005 9:24 UTC (Thu) by murrayc (guest, #23961) [Link]

That would be a problem if the issues were truly controversial. The problem here is that large groups often spend their time talking about unimportant issues, because they can. The difference between two decision is often less important than the difference between one decision and no decision.

Anyone thinking of "Highlander" yet?

Posted Oct 27, 2005 6:14 UTC (Thu) by JoeBuck (subscriber, #2330) [Link]

"There can be only one!"

FEWER = BETTER

Posted Nov 7, 2005 21:42 UTC (Mon) by superstoned (subscriber, #33164) [Link]

as gnome's interface and development seems also to be founded on 'less is
better' (features, buttons, options, colors, whatever) i think you are
right. why not reduce to 2? 1?

GNOME hits a decision point

Posted Oct 27, 2005 13:00 UTC (Thu) by gstein (guest, #3612) [Link]

The Apache Software Foundation has a nine person board. The number was chosen to be relatively large because all the Directors are volunteers. This allows a few to be missing at Board meetings, yet still allow us to reach quorum and get stuff done.

If the GNOME board has problems reaching consensus due to its size, then they need to do a bit of navel-gazing to figure out why. The Board should be very unified and be following a single direction. At the ASF, the Board typically passes *all* resolutions with a unanimous "yes" vote. We work very well together and are following the same path. Every now and then, we don't fully agree on something; in those cases, we typically table the motion or resolution until we can talk it through and reach a solid consensus.

An eleven person board *is* a bit large. The ASF is at nine, and the Python Software Foundation is at seven. I was never happy with the seven person board for the PSF, but so be it. GNOME could shrink to nine, but I wouldn't advise less than that (especially given the community size; there should be an ample population to draw Directors from). The board should also grow a very clear purpose and position which can interest people in participating; having "too few" people at board elections is a major problem. With lots of people interested, then the seat-warmers can be booted, too :-)

The decision/resposibility problem is an interesting one. We've seen the same at the ASF and only within the past few months have found ways to work through that.

I don't know if the lessons that we've learned on the ASF board will apply to GNOME, but it would be interesting to compare/contrast and see if some of the thoughts can help. Looks like I have an email to send :-)

Cheers,
Greg Stein
Chairman, Apache Software Foundation