June 17, 2006
Google bombing keeps booming
An enterprising spammer apparently has created over 5 billion pages and gotten Google to index them, in only 18 days. The pages carry Adsense (Google) ads, and apparently by spamming blogs they are getting enough cross-links to get ranked and hit in searches. The business model is to steer hapless searchers to totally useless pages and get Adsense clickthrough revenues.
The idea isn't new, but the scale is pretty astonishing. The Monetize blog article linked above includes step-by-step instructions on how to do this. A pernicious type of pollution, and it looks like it may be growing exponentially: a problem begging for an incentives-based solution!
Posted by jmm at 06:20 PM
Open access to Wikipedia less open
Growing Wikipedia Revises Its 'Anyone Can Edit' Policy - New York Times (registration required).
Wikipedia "organically" grows its content through the contributions of volunteers. Most of the work appears to be done by about 1000 core participants, but for the most part, anyone on the planet with Internet access can edit, delete or add content. Until recently...
This past winter, the media drew attention to a few glaring examples of subjective Wikipedia manipulation. One of the first publicized was a joke reference to John Seigenthaler Sr. as a participant in the assassinations of both John and Robert Kennedy. But then folks started noticing that politicians' pages were burnished, and other dubious content.
Now, in response, the informal Wikipedia management has implemented rules that put some limits on who may edit an article, and when. A small number of articles are "protected" so that no one can edit them (except the management team?) until further notice. Another group are "semi-protected" so that only users who have been registered for at least four days may edit. (Both protected and semi-protected articles are labeled in color at the top of the entry so that readers are warned about the controversy and the likelihood that some content is unreliable.)
For ICD, it is the latter category that is interesting. The "protected" articles simply are not open-access user-contributed content during the time that they are protected; they have become a traditional proprietary content vehicle. But the "semi-protected" articles are still open, with an incentive mechanism to discourage inappropriate editing. There are a couple of elements to the mechanism, and various ways in which these elements may interact with incentives:
Registration Only registered users may edit a semi-protected article. Since there is no identity authentication when an account is created, this mostly serves as a small cost imposed on users who want to edit. The pseudonymous identity can be used to build a positive reputation through repeated good contributions, but is not likely to do much to discourage bad contributions since the pseudonym can be abandoned. (I think, however, that the Wikipedia system also tracks the IP address that editors are using.)
Waiting period New pseudonyms have to wait four days before editing semi-protected articles. This cooling off period seems intended primarily to slow down "revert wars" when two opposing camps rapidly re-edit a controversial page. This kind of pollution is annoying, but probably not too dangerous since it is pretty evident to anyone who is making serious use of an article. The cooling off also may be enough of a time cost to discourage casual polluters who think it would be fun to manipulate an article on a controversial topic.
The NYT published a list of articles currently under editing restrictions.
Posted by jmm at 10:23 AM