When I read a book, it is in order to write a review for publication. Accordingly, I create an electronic document on which to take notes as I read each chapter. By the time I have read the book, I have taken fairly detailed notes, then I synthesize those notes further by writing my review. I find that this process helps be make the content of the book, as well as my own critique of it, a more permanent part of my personal knowledge base.
I do not read research papers in the same way. When I read a research paper, it is almost always because I am considering citing it in a research report that I am writing on my own work. My goal is to create a short abstract of my own that notes enough details about the method and results so that I can cite the findings. So when reading the report, I look for the hypothesis, scan the method, then jump to the Discussion to see what the essential findings were. Most good research reports summarize the major findings in the first paragraphs of the Discussion.