Quote:
Originally Posted by kguil
@YRead Currently, Marvin is very strict about the fact that they have to be identically the same book.… I'll look into this to see if I can make the matching procedure a bit less strict without causing other problems.
|
Understand it could be a difficult problem. Likely you know already. There are a few UUIDs (if really unique Ids) to choose from in an ePub: in toc.ncx, the Adept.expected.resource in html headers for commercial works with removed DRM, or some dc:identifier in the OPF. Not exactly sure how unique some of them are, but perhaps falling back to the old method if a UUID isn't found. I'd guess you might have already been planning something like that. One possible option would be somewhere being able to match up a book with a set annotations. I am a bit hesitant to suggest new features that might not be used by many, in the eventually there are just too many options, but that might be a way.
All this you likely know. Just wanted to say, as I rely more on M3, hope the new M3 annotation format is robust enough so that for future upgrades, they are fully preserved. Possibly they are already.