Yeah, it sucks.
Another is the otherwise excellent Tal's winning chess combinations, where the preceding text sometimes gives a clue, but often not. What is especially infuriating, is that the diagram somtimes starts with a Black mistake and then shows how White takes advantage.
Diagrams in some of chess books drive me crazy because they don't indicate which side is to move. My method of working through book examples is to try to understand the position by myself and only then to see the book analysis. If the side to move is not indicated I am forced to read the text first.
Some publishers are very consistent in this sense. Gambit Publications, Quality Chess – each diagram in every book has W or B or white/black triangle or square next to it. But many other publishers put it in one book, but not in the next one. I don't get this – shouldn't it be a standard in XXI century chess publishing.
Here is my Hall of Shame:
The ChessCafe Puzzle Book, Muller (published by Russell)
100 Endgames You Must Know, de la Villa (published by New in Chess)
Back to Basics – Strategy, Beim (published by Russell)
The worst thing is that those are excellent books probably belonging to chess classics. Why can't the editors do their work properly?
Does anybody else have such a problem? Am I demanding too much?