This is perplexing. If other hosted publishing tools adopted Knol's blanket no-follow approach it would break the concept of Pagerank.

Why ask for my real name, and give me a trusted status, then not trust me with publishing links?

Hopefully they'll gradually relax this policy. I understand it in principal though, if Knol didn't no-follow by default it would get flooded by spammers in about 3 seconds flat.

With luck, they'll develop a variable no-follow policy based on reputation.

Google Sites is doing the same thing as well.

Guess they want to make the place unattractive for spammers.

Didnt they say somewhere that Knols will not get additional trust just because it's a google domain? If this were true, they would not need nofollow.

Are you sure you cant take the nofollow off yourself?

I have been reading fascinating articles about constipation and heart attacks. From a samle of about 8 articles the links out from these articles look real. Not a nofollow in site. I see lots of no follows on the HP links to the authors bios. Am I missing something?

Adrian, at the moment I'm getting 2 results for [heart attack] on Knol. Both contain <meta name="robots" content="index,nofollow" /> in the header. What do you see?

http://zuosa.com/photo/mms/00/00/4D/40419884.jpg

https://www.google.com/accounts/knol20x20.gif

www.google.com/accounts/knol20x20.gif

this image did not exist.

If you look at the Aaron Wall KNOL on "SEO Basics" you can see why they nofollow. He linked back to his own site once every 100 to 200 words.

IronMal, so what's the difference to him linking back on e.g. his Blogspot blog (which is followed) or any other website out there? Linking to any place does not guarantee you good rankings *unless* Google pushes the Knol pages beyond what they deserve. Knol pages can be good or bad like any other "followed" website out there, and Google's algos need to deal with that anyway. Conversely, it seems that if an article is not good, then it hurts human visitors just as much, and should be penalized in some way... instead of sending visitors to that Google-hosted article but then not passing on linkjuice to sites hosted elsewhere.

Philipp: I reckon it has more to do with discouraging spammers under the principal of "don't bother spamming this site, it won't help you" than penalizing people who self link.

That said, I don't really see a problem with self links in this context, as long as the article they're included in is still useful / informative.

> "don't bother spamming this site, it won't help you"

I know what you mean, but that's the problem, it *will* help spammers if the article itself ranks well... because nofollow and linkjuice aside, a lot of human visitors may click the links anyway.

> ...there's no invite link so far for moderated collaboration.

Surely that's missing the point? A knol owner doesn't know the people who are going to submit suggested edits for moderation. The whole world can suggest improvements, so how could an invitation system work?

[sorry for the potential confusion in my trying to pull an important side part of this discussion into another thread now linked from the blog – http://blogoscoped.com/archive/2008-07-28-n88.html]

> A knol owner doesn't know the people who
> are going to submit suggested edits
> for moderation.

Two thoughts on this:
- I wanted to invite Tony Ruscoe to do exactly that, come to my Knol article and then add changes, but let me see the changes too before they go-live. So I knew him.
- You are right, there may be others who may want to help out. But how do you set the system to emphasize this "call for comments" kind of article? Is just setting it to "open" enough?

The most interesting thing about the method of NoFollow chosen is that it is global to the page, thus includes Googles own navigation links within the Knol pages.

In some ways, I find that a lot more honourable and fair than sites that fudge in a NoFollow just on posted links.

Of course, it is also interesting that the method chosen is older than any issues about link-juice. Its a fundamental part of the Robots Exclusion Protocol, and the META tag was originally created for those without access to the root, such as with free hosting in a subdirectory, etc.

Perhaps this isn't about link-juice at all? Or what if it is simply a method to prevent one very well respected Knol article passing too much juice to another, quite worthless one simply due to some proximity of topic or submission time?

Remember, the one unique thing about this method of preventing the following of links is that it hits THAT site itself as much as links to external resources. Somehow I think there's a reason for that.

What do you think?

The whole idea of PageRank is rather outdated, and to think that adding a "nofollow" tag is a valid solution to spam is simply stupid, those using the nofollow tag on external links are eliminating the juice of geniue links, and they are making the PageRank system obsolate.