There are many good answers here already, but I'll just add an observation on my own technique once I have decided on a book or paper to read. I like to read fast at first to get an idea of the lay of the land, so to speak, and keep on reading until I am lost due to not having absorbed previously introduced concepts. Then I backtrack and try to tease out the essential definitions before proceeding. Also, rather than reading proofs line by line I take a quick look and then try to fill in the details myself. That's time consuming, though, so it is reserved for proofs I really want to learn. This method is probably not suited for everybody, though. Everyone has to find their own method for absorbing new material.