Thursday, 21 February 2008

One Committer, One Vote?

I'm sure many folks will assume that all committers, if not created equal, are at least treated as if they were. It seems like one of those basic principles of fairness which we'd like to believe is one of Eclipse's corner stones. But, as a board member, I've learned to read rules very carefully so I can more easily subvert them, err, I mean so I can better follow them religiously. I've found it pays off to look closely at the details.


If all you keen voters out there did the same, you'd carefully read the rules governing Eclipse's Election Process. Following the various links, you'd eventually find this quote in the bylaws:
Committer Members who are employed by the same organization (including Affiliates thereof as defined in Section 6.2) shall have only one (1) collective vote with respect to election of the Committer Directors as well as any and all matters that come before the Membership At-Large, and such collective vote shall be determined by a simple majority of such Committer Members sharing the same employer as described above.
It clearly spells out that each of IBM's 200 active committers have 1/200 of a vote compared to each of the roughly 200 active individual committers (based on Nick's committer data). In other words, the individual committers have 4,000 times more influence over the outcome in the vote than does an one IBM committer. In fact, they will dominate the vote. Ouch!


To me, that hardly seems fair. Though granted, some individuals contribute a great deal and aren't even paid, so that evens the score a little. But clearly I won't be doing a lot of campaigning inside IBM! Their votes and mine are practically worthless. I'll definitely have to find green pastures, or is that bluer seas?


I wonder how the foundation reconciles the "simple majority" statement above with the single transferable voting approach? Inquiring minds want to know.


It would seem to me that single transferable voting is different from simple majority voting. I think the foundation's bylaws need review and updating.

To finish off my rant, I'd like to say something bad about each of the other candidates to ensure my own election victory. Unfortunately, I can only think of good things to say about all them. What fun is that? The best I can do is suggest you follow me. I'll make sure we end up somewhere very cool.


I just hope no one finds those embarrassing pictures of me and uses them to tarnish my reputation...

14 comments:

Chris Aniszczyk (zx) said...

How much to keep to quiet ;)?

Ed Merks said...

Will one free beer do the trick? If not, will ugly threats suffice? After all, people who live in glass houses shouldn't throw bricks.

Ian Skerrett said...

I think the intention of this voting scheme was to ensure one company couldn't dominate the committer voting. My understanding is that a lot of thought was given to ensure Eclipse was vendor neutral.

The flip side is that no one can accuse IBM of stacking the Board with committer reps. The IBM committer reps were elected due to the respect and contribution they make to the community.

Ed Merks said...

I'm sure much thought was given to the voting schema and obviously IBM itself was instrumental in helping to draft the rules. But I think with 20/20 hindsight, the net result is that committers for the strategic members have little impact on the final vote and that committers don't all have equal say over who their representatives are.

Of course no one can accuse IBM of stacking the board. In fact, only five people ran for five positions so it wasn't really much of an election at all last year; it didn't ultimately matter whether anyone voted nor how they voted. This year will be better. Perhaps that's one of the successes the current set of committer representatives can claim, i.e., improved awareness of the role and what can be accomplished with it...

David said...

Ed,

I have to say that I like the current format, as it does what it was intended to do, not give any one company undue influence into the process. Yes, I agree that if you work for IBM or another large strategic partner that is putting large amounts of resources in place that the vote is dilluted. But again, I think for good reason, to try and minimize the influence one large strategic member could have over the entire organization.

The other way to do it, is to say that one one member or one committer from each organization could vote regardless of how many committers they have in place.

Depending on the eclipse project, the strategic committers already have a lot of influence of the direction of a project and what items go into the official releases. The vote as it is set now helps to counter balance that power a bit.

Ed Merks said...

David, I think Doug agrees with you, but I don't. You might want to look at his blog and my comments there.

To address your specific point, the committer representatives aren't on the board to give more influence to any one company, they're there to provide the committers themselves with direct influence independent of any company.

I'd rather the voting rules be color blind rather than painting me blue. And I'd rather the election be truly reflective of the committer population rather than reflective of the organizational affiliations. The board itself and the elected add-in provider members are already there to represent the organizations...

David said...

Ed, yep I understand the issue and it's a tough distinction to draw the line at due to the various ways that committers are added to a the eclipse projects.

If every committer had to go through the time tested, rigorous process that individual committers had to go through, then I would agree that one committer, one vote. But the fact of the matter is that a Strategic Partner can contribute many committers to a project that may not have had to go through the rigors that an individual committer needs to go through. So, in that case, whether intentionally or not, a group of committers from one organization could influence an committer vote in their favor. Not saying that has happened or will happen, it's just a potential situation that the current rules guard against.

Maybe there needs to be a clear distinction of where the rule applies and where it doesn't, but that is going to be hard to agree upon (not impossible, but difficult).

Darin Swanson said...

This is great and healthy discussion..thanks Ed! I have gathered my random thoughts here.

Doug Schaefer said...

I have no problem with one committer, one vote, if the playing field was level. Currently it is not. I know all the committers at Wind River will vote for me because they feel an a kinship with a fellow employee. It would make sense that IBM committers would do the same. Do the math...

Doug Schaefer said...

BTW, I love this debate. Having been a fellow employee I understand things from your side of the fence. I am certainly sympathetic. But someone needs to represent the other side of the fence :)

Kim Moir said...

Ed, thank you for bringing up this issue. It has always bothered me that the votes are aggregated from member companies. We are all committers, and deserve to have our vote count. Others may see just see blue, but I see a rainbow of perspectives on committer issues. I think it's a an oversimplication to expect that people will always vote for representatives that work at their company.

Donald Smith said...
This comment has been removed by the author.
Donald Smith said...

(repost, too many typos :)

It may be an oversimplification, but what I'm actually fighting is perception.

It's my job to talk to new membership prospects, network at trade shows, get involved in competitive discussions (from a business perspective, Wayner does the technology competitive stuff), etc.

I'm here to say loud and clear -- the perception of many is that IBM is in control of the Eclipse Foundation. It's a false perception, and definitely on the downswing over the past few years, but still very much there. It's only after we review the Governance model (including rules for voting board representatives) and give lots of case studies of the diversity of our Ecosystem that our point is made.

I worry that if we change the voting mechanism we are sending a message that is counter to our promotion of open governance and no single point of control. It's not impossible to overcome, but it does give our nay-sayers ammo to spread FUD with.

Does anyone have suggestions on a voting process that empowers all developers while promoting diversity?

I do have some ideas, but they're too wickedly complicated for a blog comment...

BTW - IF there are changes to the voting process, how about only allowing votes to committers who have made at least X committs to their project in the past year? I.e., only active committers vote.

- Don

Ed Merks said...

Donald, I'll keep reminding folks that the committer representatives are on the board to bring forward the perspective of all committers, not to reiterate the perspective of the organizations for which they work.

IBM's one vote on the board is minuscule given a board of roughly thirty members. Anyone just needs to do the math, so I don't imagine it's hard to point that out to the naysayers. I have a hard time imagining that this nuance of the committer representative voting rules even comes up in discussions.

All that being said, he who does the actual work has the real control and therein lies the more substantial issue. Promoting diversity must focus on improving the diversity of the committer population itself, as Rich and I have done in the Modeling project, not on voting rules that create a large class of second class citizens.

It should be clear that the diversity of committer representation on the board will not buy us the kind of diversity we actually need at Eclipse, and in fact will not even help with that issue. So I think we're focused on the wrong issue and using tactics that run counter to basic principles of equality and fairness. One voter having one vote is a fundamental principle in most democracies. Can we point at one where that's not the case?

Allowing only active committers to vote would be good, but how to measure activity would be a problem. Dave Steinberg, for example, has been slaving on the EMF book for a very long time and will appear to be quite inactive in CVS...

I'd actually be a bit surprised if the strategic members consider the current committer representative voting rules to be a good thing. As a block, the strategic members have only 20 votes while any individual committer has as much voting power as any one strategic member that might be funding the salaries of dozens if not hundreds of committers. So I'm not sure they'll perceive the current rules as being all that fair either.

Whether I'm elected or not, I'll continue to focus on diversifying my projects and on ensuring that Eclipse's processes are fair and sensible...