The other day, I was hypothesizing how one could provide incentives for scientists to comment on papers and blogs and to blog themselves. My solution was that there needed to exist a service which would allow researchers to somehow build a reputation by aggregating their papers, comments, blog posts, etc. As I learned from Michael Nielsen on FriendFeed today, there is already a lot of experience about how to set up Reputation Systems! Joshua Porter has two great articles on bokardo.com on "Social Design Patterns for Reputation Systems". When I read the introductory remarks, it occurred to me that of course this sort of system already exists in a lot of places and that a whole bunch of very clever people must've already spent long hours thinking about similar problems for their communities:
The articles link to Yahoo’s social design patterns for reputation systems, but the main content is an interview with Bryce Glass, an interaction designer at Yahoo! Glass explains the ideas and concepts behind Reputation Systems and the different way of implementing the right pattern for your community.
The content of the interview is especially pertinent with regard to a recent analysis by Euan at Nature's Nascent which showed that only 2% of BioMedCentral papers attract comments.

For me, this basically means that all the expertise and technical prerequisites are there to bring the scientific community into the 21st century. The advantages of the new system need to be succinctly summarized and widely publicized at the same time as the current system's disadvantages and idiosyncracies need to be pointed out and publicized along with the new proposal. And because criticizing is always easier than advertising, I'll start by summarizing why Thomson's Bibliographic Impact Factor (BIF) is dead.

UPDATE: Actually, already last month, Maxine Clarke over at Nature Networks posted a question where these two articles provide all the answers: Better metrics for an individual's "value"