Closed marisademeglio closed 3 years ago
My gut feeling is that, at least in V1, we should make it simple to cover 70-80% of the use cases. Which probably rules out (3) for me.
I agree. There's no use case that would require the complexity of (3). Personally, I like (4) because it's generic and also simple. It would allow having a series of text-only playback objects, which is currently not required by any of our use cases; however, I can envision an interesting scenario where there is no audio but you're manually controlling highlight progression (e.g. "next paragraph") - perhaps to help readers focus on a small portion of text at once.
If we see this as a MO evolution, replacing epub:textref
properly would require (3), so that's what I described in the new draft
There are no restrictions on parallel media objects: https://w3c.github.io/sync-media-pub/sync-media.html#media-objects
Playback behavior is defined as "wait for the longest timed media object to finish" aka SMIL endsync. TBD: do we need to define what the other objects do during this time, e.g. fill=freeze. I believe @larscwallin has some opinions ;)
We often show how a playback object may contain text and audio, e.g.:
But we need to define what is required and optional here, i.e. how many
text
andaudio
properties are allowed or required.Options:
text
OR justaudio
ORtext
+audio
together. There is currently no defined playback behavior for a standalonetext
property with no associated timing information.