This is often known as "collaborative filtering", and pops up in systems like NoCeM and GroupLens. What's cool is that you don't need transitive trust or even poster reputations (anonymity without so much vandalism!). Just give the right reviewers the reputation "good/bad judgment about which articles are worth reading"--and you can find them by reviewing some articles yourself and measuring similarity of answers. If reviewers aren't willing to highly rate articles they disagree with there's a danger of shutting out unpopular ideas (a common complaint about /. moderators), and you need a critical mass of people willing to rate most of a forum to get started (and it can't be much more than an extra keystroke or two or they won't bother). <URL: http://www.cm.org/faq.html > <URL: http://www.cs.umn.edu/Research/GroupLens/ >