It’s midterms, and I probably shouldn’t be blogging (instead, I should be focusing on studying up on information retrieval, search algorithms, and the vocabulary problem), but there’s been so much chatter out there today, I can’t resist adding just a tiny bit more noise.
Mainly, I have a question in response to Siva’s question: what is the gold standard for the kind of search he describes? Is it a library catalog? Because I tried a few naive searches of the type he describes on Mirlyn, the U of M’s catalog, and didn’t get a heap load better results than he did on Google. For example, a “Words Anywhere” search on “copyright culture,” while it does return Copyrights and Copywrongs, does not return Free Culture. I would submit that similar things would happen given any search interface that relies wholly on unstructured keyword searching.
A central value added in a library catalog is the possibility of title searches, author searches, etc, but the type of user who knows how to use those capabilities would probably be the same type who would think to use Google’s “Advanced Book Search” feature, using which a search for “Magic Kingdom” in “Title” brings up Cory Doctorow’s book both first and second.
I also agree with a comment on Siva’s post that asserts that the major value of Google’s interface will be for browsing, not for known-item searching. The serendipity of finding things you don’t know exist — like all those books Siva popped up with his “Magic Kingdom” search — to me, is a value in itself.