I started posting into this category on my blog in March, when I was thinking about a narrow phenomenon: how Freebase and services like it would change the way we find information using search engines. A search for “employees of Plaxo” could return structured data from one or more semantic web/user-contributed metadata sites, before listing the traditional keyword results.
But I also had a related idea about using influence data to build better music and book recommendations. Knowing the author’s influences and peers, your favorite authors, your friends’ favorite authors, scanning social sites for correlational data, and being able to combine these factors to produce a personalized recommendation would be invaluable.
It now looks like the OpenSocial platform will make possible this second, socially-enhanced search. Greg Linden regularly links to papers by people at Google working on news personalization and recommendation, and the flurry of OpenSocial speculation hints at the same thing:
Mark Canter: If four people type in the same search and they have four different sets of friends, wouldn’t the results be different based on what your friends have been searching for? Google maps, reader, gmail, Youtube… Google is primed and ready to take in the social graph and do something with it.
