I pretty much agree with Aaron Haspel’s rant about footnotes and endnotes. I think Deb Nolan and I did a good job of these in the Notes section of Teaching Statistics: A Bag of Tricks. But we should have referred to page numbers as well as section numbers.
In all my books I’ve been careful to put the references at the ends of chapters or at the end of the book, rather than threaded through the text. I don’t like in-text references (for example, “We consider the xyz model (Jones, 1994)”) because they seem to me to be a way of passing the buck. I want everything in my book to be something that I believe and stand behind. But then at the end of the chapter or book, I do want to credit where I got the ideas from.
The absolute worst reference system I've seen was Eric Beinhocker's Origin of Wealth. First, it involved endnotes, which use the MLA reference. So if I wanted to see the source of a sentence I had to look up the endnote to find "Smith (1989)" then turn back to the bibliography to find that specific reference. That involves no less than three finger bookmarks.
My thoughts from last year on this topic.