[NiLuJe]

@Ripplinger: I tend to agree with you on that, FWIW.

My main "issue" with that approach is with older books imported before I actually thought this through (which was when working on the Kindle Collections plugin, where I picked up a dedicated trick: I prefix the "main" tag (i.e., genre) with a specific character (e.g., '+'). The plugin can then detect that via a user-specified regex to pick the "right" tag/tags from the mess):
Those are full of crappy tags that I'm way too lazy to cleanup ^^.

(My current radical answer to that fact is to not use collections at all ^^).