Friday, September 16, 2005

Warning: nerd post

The general idea behind Google's ranking (and the other major search players) is that a page's rank is directly proportional how much time someone who randomly clicks links would spend on that page.

It's called the random surfer model - if I click around and spend 1 second on every page, but spend 5000 seconds on yahoo.com and 5 seconds on stupidpage.com, then Yahoo's rank is 1000 times greater than stupidpage's.

But all of this is neglecting how people use the web now. Who actually surfs? Is there a better way to rank pages, having to do with how emails, IMs, and away messages fly around for faddish cult favorites? What about with the fact that searching is how I get to most pages, not browsing?

Don't get me wrong, PageRank works pretty well, but I'm starting to get annoyed with some of the automated properties of it: like how the top result on lyrics search invariably starts the JVM and tries to do a java exploit

Okay, so until next time, keep sending me your questions, and I will keep making fun of your punctuation and spelling. I mean, answer them.

No comments: