Greg also provides a summary and a bit of analysis. This paper was presented about a year ago at the Beyond Personalization 2005 workshop. If this topic is of interest, plenty more papers and position statements here along with the results of the workshop (worth reviewing).
Btw, the site where all of this content is found is hosted by GroupLens, a research group at the University of Minnesota that has been doing in collaborative filtering since 1992. If you're a movie buff, take a look at MovieLens, a publicly accessible and free movie recommendation service that has also been online for several years.