Google Blogoscoped

Monday, January 22, 2007

Wikipedia Nofollows Links

Wikipedia now adds the “nofollow” attribute to all external links. The reason given for this is to battle spam links, like those of blackhat search engine optimizers. The change is already active, as you can see by checking outgoing links on Wikipedia’s articles. (In German Wikipedia, this change was already activated for a longer time.)

Such a change in Wikipedia, with its millions of pages – many of which rank excellent in Google and have a high PageRank – has a potentially strong impact on Google search results. Google relies on links to determine its result rankings, and thus the huge amount of outgoing links on Wikipedia do their share in influencing that.

Originally, “nofollow” was aimed at comment spam of blogs and such. It then entered the realms of online advertising links. It now covers wikis as well. (The attribute’s name, “nofollow”, is unclear on whether any of these applications are correct; no-follow doesn’t describe what the link is, but rather, what searchbots are supposed to do with it – this is also why we have endless discussions about where and how to implement nofollow.) Back in the beginning, Google – co-inventor of the attribute, along with Yahoo and others – announced that nofollow ought to be used ...

... anywhere that users can add links by themselves, including within comments, trackbacks, and referrer lists. Comment areas receive the most attention, but securing every location where someone can add a link is the way to keep spammers at bay.

Clearly, wikis fit this description, as everyone can add their links (ad links, by the way, don’t fit this description, as ads are added by webmasters, not users). However, the idea of wikis is also that they have a self-healing capacity; every link posted by someone can be easily removed by someone else if it appears to be spam. In fact it’s through exactly this self-healing capacity that Wikipedia justifies its concept and the usefulness of their articles. Someone can post a falsehood – and there are many incentives for certain private persons or companies to do so – and someone else can replace that falsehood with a fact. You may not believe in this self-healing capacity, but surely, if you do you assume it applies to links as well; a link can be just as valuable or invaluable as every other unlinked part of a Wikipedia page.

What happens as a consequence, in my opinion, is that Wikipedia gets valuable backlinks from all over the web, in huge quantity, and of huge importance – normal links, not “nofollow” links; this is what makes Wikipedia rank so well – but as of now, they’re not giving any of this back. The problem of Wikipedia link spam is real, but the solution to this spam problem may introduce an even bigger problem: Wikipedia has become a website that takes from the communities but doesn’t give back, skewing web etiquette as well as tools that work on this etiquette (like search engines, which analyze the web’s link structure). That’s why I find Wikipedia’s move very disappointing.

“Nofollow” was always a highly two-sided initiative. It has pros and cons, and until now, it was hard to tell if the cons were ever able to outweigh the pros of this invention. Now we have another example of the can of worms Google and others opened when they introduced the attribute.

I predict some people will now, in return, stop linking to Wikipedia, or “nofollow” their links to Wikipedia (following the argument that if they don’t trust their own system, we shouldn’t either, and also following social etiquette – returning a disfavor, so to speak).

Wikipedia would fare much better if they acknowledge the wiki spirit of communities working towards improving articles over time. And there’s a trivial way to ensure “nofollow” becomes sensitive to time. I implemented it in this blog’s comments: any fresh link that is posted receives a “nofollow” attribute, but only for a couple days, upon which the “nofollow” is automatically removed... turning this into a normal link. The reasoning behind this compromise is simple – searchbots will not value spam links here (giving spammers less incentive to post them, too), but any link that is not removed by one of the comment moderators over some short time must be a “trusted” link. The comments still respect the nature of links, and the nature of a community who adds value with outgoing links.

I’m sure we don’t catch every spam link here with the “fading nofollow” approach, and I’m sure neither would Wikipedia when using the methodology; but 90% working solutions are exactly what Wikipedia is all about. Try to make Wikipedia 100% perfect, and you’ll need to remove the “edit” button... a point at which the system becomes very far from perfect, very soon.

[Thanks Haochi in the forum!]

Advertisement

 
Blog  |  Forum     more >> Archive | Feed | Google's blogs | About
Advertisement

 

This site unofficially covers Google™ and more with some rights reserved. Join our forum!