Monday, June 12, 2006

Link aggregation and alumni organizations: avoiding the fraud and diversity problems

I was thinking this morning of the problems that affect all intellectual products that emerge from community efforts -- from link aggregation web sites to Wikipedia to Google's search data.

All suffer from two problems: fraud and diversity. They have something in common with each other (and with biology of course).

Fraud is the creation of material that pretends one agenda, but serves another. Spam blogs, or splogs, now infest Technorati. They are a degraded product. Authentication and reputation management are the technical approaches to this problem.

Diversity is trickier. One man's meat and all that. Darwinists and anti-Darwinists, scientologists and rationalists, alternative medicine and science-based medicine, jihadists and humanists -- we all have rather different interests.

Sometimes I do want to see what the jihadists or the scientologists are saying -- but not usually. How can one approach this problem?

Google tackled it by creating products like Google Scholar, which rely on sources that are authenticated and on age-old filtering and reputation management systems. Google scholar is great, but it doesn't help me figure out what iPod clock radio to buy.

Amazon now gives more points to their authenticated reviews and attempts to manage reviewer reputations. Their system, however, has been extensively corrupted by vendors (IMHO). It is still true that negative reviews, when cross-checked by positive reviews for refutations, are very valuable (though I'm sure rivals are now faking negative reviews as well).

I think there's another avenue to explore -- alumni organizations. This will vary by institution. I'm an alumnus of McGill University, The University of Minnesota, the University of Quebec at Chicoutimi (sort of :-), Williams College, Michigan State University, a few other places, the Watson Foundation, and Caltech. All but the last couple are too diverse to provide any useful filtering and validation. Caltech, an institution that's at least 99% rationalist, and (to a lesser extent), the Watson Foundation, would be very interesting.

I wonder if I can persuade the Caltech alumni organization to start doing some authenticated implementation of one of those link aggregation web sites. Maybe I could sell it as a marketing tool, and the Ad Word revenue might help a few students ....


Andy Shaindlin said...

Thanks for this idea. We're educating ourselves right now about "link aggregation communities" and will look at how this might work and what the benefit might be.

Andy Shaindlin
Executive Director
Caltech Alumni Association

John said...

I think Caltech gets points for posting their reply to the blog! Good for my alma mater.