I just ran into an interesting article about a project called Reference Extract. The idea is to build a search engine that provides more relevant hits for search terms and it tries to accomplish this by using real live librarians.
Search relevancy is easily illustrated every time you search for something. Google indexes so many pages that when a user searches for a term like “bostelle UWT” because they want my work phone number, the correct hit is all the way at the bottom of the page. Information experts know that most users will actually try the search again rather than even bother looking past the first two or so searches.
Thus relevancy is crucial to making a search engine work and Mike’s group wants to make search engines work better.
The problem is that Google has “pretty good” relevancy, is easy to use, widely accepted, and widely trusted (rightly or wrongly). I think what these librarians are trying to do is noble and encourage them whole-heartedly in the endeavor.
But that’s a huge mountain to climb.