"nofollow" is a link attribute which tells a search engine that it should not follow the link, i.e., the link will not get PageRank or any other benefit from being on the page. It's primarily intended to stop link-spammers from taking advantage of comments or open forums. However, it has other uses, such as PageRank "hoarding", a kind of no-PageRank-for-you! statement.
There's been a longstanding debate as to whether Wikipedia should add "nofollow" to non-Wikipedia links. Pro: Wikipedia is tremendous spam-bait. Con: Good sites deserve the search benefit from one of the most popular sites on the net.
The previous community decision was:
As of May 22, 2006, following this discussion, rel="nofollow" is now enabled on non-article pages (i.e. pages outside the main namespace) on the English Wikipedia, but remains disabled for links in articles. Brion has said that it is his "intention to enable nofollow everywhere in the long run (though this might end up being in more limited form, for instance allowing some whitelisting or other verification process)."
That earlier decision has now been overridden by "Jimbomancy"
Having been requested by Jimmy to do so, and having seen a fun rumor of a "search engine optimization world championship" contest targeting [Wikipedia], I've gone ahead and switched rel="nofollow" back onto URLs in en.wikipedia.org's article namespace.
Note one implication to draw from this:
WIKIPEDIA IS NOT AN ANARCHY! THERE IS SOMEBODY "IN CHARGE"!
Let me be clear - in many cases, I think Jimmy Wales' decisions are right, and in fact a necessarily corrective to the impulses of crowds. But they sure are top-down CEO-type actions. The propaganda of Wikipedia should be proven transparently false every time one of these events happens. Wikipedia's social organization should be familiar to anyone who has seen a bureaucracy where the low-level administrators aren't accountable much to anyone else, and the bulk of the work is done by volunteers. Getting this often-dysfunctional setup to work is an achievement, no doubt about that. But it's very limited in terms of how well it scales and how much it's a model.By Seth Finkelstein | posted in wikipedia | on January 21, 2007 01:00 PM (Infothought permalink)