Relevance for tweets / status updates is pretty simple: Newer = more relevant.

For example: http://www.google.com/search?q=san+francisco+halloween gives me a bunch of halloween stores

http://twitter.com/#search?q=san%20francisco%20halloween gives me a bunch of events (much more useful for me!)

If those sort of tweets get into search results simply based on "newest", you can be sure it'll be used extensively by marketers trying to spam their way into search listings.

FWIW, Twitter seems to be on top of the arms race right now. A little while back I was getting 3-4 spam followers/@-replies a day; over the past few weeks, though, I've had exactly two.

I am getting these almost every day now. Annoying at first, now I am considering using some other service. Twitter is too popular and popular services eventually become shit.

> I'm not sure I want search results to be littered with tweets.

Granted. I hope they find balance, and imagine they'll AB test to see what users want.

> How do you apply page rank/relevance etc to 140 chars?

Some combination of freshness, quantity of retweets, pagerank of any linked pages, bit.ly stats on links, historic tweet quality of the user, similarity to trending topics, follower count of user, follower/followee ratio, incoming links to the page from blogs, etc, etc, etc.

retweets are one way to quantify relevance. "real people" are not going to retweet something unless they find it meaningful.

I like the idea, but what's going to stop a bot from re-tweeting?

If a lot of people have retweeted you, you are ranked higher on the list, which means that your retweet is ranked higher than retweet from a bot.

Yes but nothing stops a botter from running thousands of bots re-tweeting the main spambot's tweets.

Disagree. As I said a week or so ago. Retweeting is often the same as an email asking you to forward it to 100 friends to get good luck. I don't think that makes for very good measurement of what is important and what isn't.

That is what the engineers get paid to do:)