Fair enough. I take a different approach in my interviews: try to gauge how useful the person will be to the team. We have internet access, and a library with CLRS if someone needs to implement a bipartite graph test. It matters more to me that they'd know why they'd need one.
We use those kinds of approaches too. I never said the interview consists solely of closed book tests, they're just one thing I find useful to help gauge the suitability of an interviewee.
[EDIT] The thing I'm most looking out for is how a candidate answers when they don't know the right answer.
