graywolf - 3:16 am on Apr 2, 2005 (gmt 0)

As we know Yahoo keeps tracks of searches and publishes breakout or noteworthy increases and decreases weekly. We also know google tracks and publishes less data in a less transparent format in the zeitgeist.

If you wanted to detect unnatural link bloom, one way would be to look for an increase in links without an increase in search volume. The greater the difference the more spam like it looks. This would also help them know when to deliver fresher documents as opposed to stale ones.

<tin foil hat theory>You could break searches down into categories (think Dewey decimal like system) and establish more accurate baseline measurements for categories. Searches for people like Britney or Paris would naturally be spiky. Searches for euclidean geometry would be relatively flat</tin foil hat theory>

I think recips, or more specifically recip pages are going to be "so 5 minutes ago". For example if I put up some amazing content on my website and got a link from the New York Times, would I be devaluing it by putting up a link back to the Times website on my home page? Lets face it from an spiders point of view a recip page is pretty easy to pick out. The format and link to content ratio is probably fairly consistent across different sites on the web. However if the links were placed inside articles or even relevant blocks of copy it would look less like a reciprical link directory.

My two drachmas.