Thursday, April 26, 2007 friends

Thinking again. One of the things Steve Poland is going after is the concept of webothlike:

Basically the idea of is to connect like-minded people — people that have the same interests. You’ll go to the website and answer questions, as many as you want. The more you answer, the more profile data we have on you — and the more we can contrast you with others in our system, and match you up with them

I think rather than asking questions it would be better to try a recommendation style algorithm on your links, or bookmarks, etc. so that you could get a match score for how similar you are to other people. Basically you could submit all of your links and get a quick score for who else is tagging the same things as you. You could take a similar approach with OPML comparisons to find who is reading the same things as you. On the one hand, the output of this would be suggestions about what to read- and really, who doesn't have too much to read already? On the other hand, a cooler output suggested by WeBothLike would be to find people to scheme with.

Anyway, I have a feeling someone has already done this somewhere, but it's not part of the main application, and it probably should be.