You’re pretty much on the ball with your analysis of chuquet. Just a couple of quick notes: the link ranking is based on an algorithm of the number of links and how recent those links are. The math is being tuned over time to get it ‘right’.

I guess we need to clarify that on the site. I could show the ’score’ that we use for ranking, but that’d probably just make people wonder what the algorithm was doing. It’s not linear!

The tag cloud is based on an analysis of anything in a blog post flagged as a ‘tag’. We then count and accumulate these, and link the tag back into chuquet as a search.

Our plans for personalisation are very much based around OPML, but the challenge is to *reduce* the result set rather than increase it! Or at least allow the user to tune how many results they get and of what relevance.

And finally: thanks for the post! I appreciate it.