This was originally a longer post about an interesting couple of graphcis posted by Matt McAlister in What will be the next PageRank? but Firefox crashed out on me just as I was about to post, so I lost everything...
Anyway, the gist of that now lost post was picking up on the question of what could come after PageRank as the number of failed searches grows in the face of an increasing volume of information and a worsening useful:useless information ratio (by the by, FF crashed just as I google 'PageRank algorithm' so I could give a link to the original paper...if you want the link, go find it;-)
Matt comments:
The click is much more potent than the existence of a link. Even more potent than clicks are tags, ratings, comments and emailed URLs. A hyperlink is still a vote, but seeing some form of human action gives me much more confidence that a source has value.
Picking up on the first part of that quote put me in mind of the Attention Trust. If we record all our interdomain clicks, anonymise and aggregate this information, then are we a step along the road to a ClickRank inspired search engine?
NB I also wonder how the various people powered search engines (PeopleRank?) are gonna fare this year? Perhaps information about sites included in user-specified searchrolls or OpenSearch profiles can also be used to provide intelligence about the sites people really rate as information sources.
This is exactly the sort of question I'd be asking if I was a library 2.0 person, I think...
Posted by ajh59 at February 19, 2006 10:40 PM