On Feb 6, 2005, at 6:21 AM, Magnus Manske wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Once upon a time :-) I wrote a wikipedia mode that would display wiki pages (no edit), and when hitting a non-existing link as a logged-in user, one would get a framed (evil!) page with the matching wikipedia page, a small header showing an "import" button. "Import" would copy the wikipedia text to that wiki. It even did automatic image imports.
What is already happening is that Wikipedia is moving to a system of experts, only they are experts not in any particular subject, but in the ability to get wikiprivs such as admin status, telling editors to cool down and forcing some kind of consensus. The reasons for this are simple to outine. First, most people don't apply abstract rules such as NPOV, no original research or encyclopediac. They apply simple social rules: namely three reverts and don't say anything bad about someone. Which means that many revert wars really start on talk pages open with crypto-insults - the "fourth person derogatory" - attacks on "those people who are ignorant about my POV".
The first wave response then isn't to negotiate - because telling an idiotic jerk who has a thinly sourced POV to push that he is an idiotic jerk with a thinly sourced POV counts "against" the editor who said it. Since we've outlawed truth, the level of wikistress rises, and eventually one side or the other tries the simple mobocracy road. Why get someone who disagrees with you to agree with you, when all you really need to do is get someone who agrees with you to agree with you and vote the other person down. Since the sample size is small, and the methodology one of agreement, there's no assurance of converging on anything other than polarization.
At some point a revert war breaks out, because one side thinks it has enough of a numerical advantage, and someone with admin powers comes in, issues a bunch of warnings to everyone, the way a principle blames every kid in a school fight, and the consensus is forced, sometimes with the admin taking over the role of chairing the article. Depending on the admin, this can slant the article significantly or not.
This road is the path of least resistance because at almost every step of the way, the dominant strategy of equal "normal" editors is to betray discussion. Since mechanisms such as mediation, RFC and arbitrarion are slow, and generally only draw admin attention, it will continue along these lines.
My suggestion on this matter would be to speed this process. First by making the 3 revert rule apply to content not individuals. No more than three reverts of a particular section, regardless of how many editors on each side. We create a category "reverting articles", which the person who reverts the article for a second time in a particular direction must add as a tag - short and sweet - and then a category "revert limit" articles. People interested in preventing revert wars, of which there are a number, can then simply read the list, and get involved. It won't help if they are simply members of a particular POV, since the limit is by number.
We can even create subcategories for why there was a reversion. This makes disputes less visible, and means that formal mechanisms are for when there is some clear conflict which requires larger comment. It is only slightly slower than simple reversion (one must type in a tag), but much faster than RFC.