hmeine
commented
11 years ago I have now implemented three ways to recognize slides:
- by headline (early versions)
- by new content (later versions, also not satisfactory)
- by latex-beamer \labels (works best for me)
Currently, a slide is defined as a stack of frames (PDF pages) with the same headline. Given a better model of the frame contents (cf. issues #3 and #2), one could e.g. start a new slide as soon as some content disappears.
The main goal would be to optimize the miniatures in the overview (basically showing all available content). Given the better transition system (issue #3), the presentation itself would not be affected anymore by different slide definitions.