I was hoping to get some of your thoughts on best practices for TR1 usage in boost libraries (maybe there is already a recommendation that I haven't seen yet). I'd like for some of the libraries I'm working on in the sandbox to use the TR1 namespace wherever possible, since this seems more future proof than using the boost namespace. I understand how to use the TR1 namespace in boost so that it uses the TR1 implementation of the compiler provider if it exists, or the version of boost if it doesn't. For libraries in boost, I don't see a really "clean" way to do this as it seems it would require the user to add an additional include path beyond just the standard boost path (just like you do if you use it outside) to use my library. Is anyone trying to code their boost library to the TR1 namespace? How are you doing it? Thanks, Jared