On election timing

I blogged a few weeks ago about some issues I saw with the process for the Wikimedia Board election that was held recently. The results, as I am sure most people are aware by now, were released on 12 July, with Erik Möller, Kathleen Walsh and Frieda Brioschi being elected to serve two year terms.

One seemingly obvious observation that seems to have been lost so far amidst the larger controversies is that the new members of the Board were appointed some time after the date when their terms were supposed to commence, which was 1 July. Either the seats of the three outgoing members were left vacant for twelve days, or else they were occupied for twelve days beyond the time that they were supposed to be. Either alternative points to a serious failure with timing.

This is representative of almost all the problems with the election, which I attribute primarily to how little time there was for the Steering Committee to organise these elections; the Committee only being put in place at the beginning of the last week in May.

A number of problems with the election were caused or at least exacerbated by the haste with which the election had to be conducted. As an example, the amount of translation achieved in the 2007 election was less than that which has been achieved previously. For the 2004 election, the primary information page for the election on Meta was translated into 32 languages, and the 2005 election page was translated into 35 languages. In 2006 this dropped to 26, and in 2007 the main election page was translated into just 21 languages. I suspect that this was at least partly due to the fact that the master version of the election description page was only released a little over an hour and a half before candidate entry began. This is perhaps the most obvious problem with the election that could be lessened by allocating more time, though there are others.

Given that we know exactly when elections will be due (since we know when members' terms conclude) it should be entirely possible to plan when certain key events in the election process need to occur, and indeed to prepare for many parts of the election in advance.

I have put together some more thoughts in my userspace on Meta, along with a suggested timeline for when the election process should begin. Any suggestions as to improving this suggested timeline (particularly input on how long should be allocated to each phase of the election) are welcome.

Election process

I feel dirty saying this, but: process is sometimes important, at least when it comes to elections to the Board of Trustees of the Wikimedia Foundation.

With all due respect to the members of the Board Election Steering Committee, the process for this year's elections leaves much to be desired. It is likely that many of the problems are a consequence of the committee only being put in place at the beginning of the last week of May, just over a fortnight before the election process was due to commence.

The most glaring issue is that major decisions about how the election will be conducted are still being made even after the process has started. For example, at the time of writing, the nomination and endorsement periods have closed, and voting will commence in two and a half days, yet neither the external third-party who will be conducting the election nor the software to be used for voting have been identified yet (or if they have, there has been no public announcement). While it is pleasing that the election will be conducted externally, and will not use an on-wiki method for voting, it is worrying that these fundamental issues have not been worked out in advance.

My other concerns are structural issues with the way that the elections have been organised.

While it's entirely appropriate to have participation requirements for voters (so that the franchise is extended only to those who have some stake in the Wikimedia community) it baffles me why we have participation requirements for candidates, particularly given the requirement (introduced this year) that candidates need a certain number of endorsements from members of the community eligible to vote in order to proceed to the voting phase.

An endorsement system (with appropriate parameters) should work fine in eliminating candidates unsuitable for whatever reason from being involved in the important voting phase, but having additional participation requirements can only serve to restrict the pool of candidates. There are all sorts of people out there who might be great to have on the Board of Trustees, such as people with experience in other non-profit organisations, people involved in other free content projects and so forth.

There are other, smaller problems with the current endorsement system.

Firstly, the just-closed nomination period ran concurrently with the period for endorsements. Several people observed on the Foundation mailing list that this effectively brings forward the closure of the nomination period, and it does so by an unspecified amount, since people would have to guess how long they would need to attract the requisite endorsements. effe iets anders suggested that it should be entirely possible for prospective candidates to line up their endorsements before nominating (and this also seems to be the conclusion that the Steering Committee reached, judging by this comment), and it should, but this for me raises a question about the objective of endorsements in the first place, which is supposedly for the community to assess whether candidates are worthy of serious consideration or not. I would rather have candidates putting their efforts into preparing their platforms, writing statements and answering questions, rather than shilling for endorsements.

Secondly, while the required number of endorsements was twelve, endorsement pages were not closed at this point, and endorsements continued to be listed, resulting in some candidates gathering very large numbers of endorsements. Some users expressed concerns that this was effectively a second voting period, distracting from the real vote ("a public one and a private one", as Walter Vermeir put it). It also runs the risk of making candidates who nominate early appear to have more support than candidates who enter later in the nomination period. If the objective of endorsements is only to eliminate the unsuitable candidates, then I would think it sensible that endorsements for a candidate be closed once the required number is reached. It may be preferable to have endorsers send their endorsements directly to the Steering Committee, perhaps by email or perhaps by an on-wiki interface, so as to prevent the perception of a popularity contest.

It seems to me that many of these issues could have been avoided had the Steering Committee been formed with adequate time to prepare for the election, and given that we are now more or less on a fixed cycle of elections (with certain positions up for election every two years) there is ample opportunity for planning to occur. I also feel that, given that some of the problems I have mentioned have been identified by other people also, that the Steering Committee could have opened up some of their decision-making processes to community input. Again, this was probably prevented by the short time that they had in which to work.

I would suggest that a timetable be laid down in the near future for the June 2008 elections, and that discussion on next year's format be opened now so that there is opportunity to comment while the issues around the present election are fresh in people's minds.

Now it's time for a wash.