– DRAFT –
Publishing Maintenance Working Group Telco

Adding synced media for incubation to the new charter

wendyreid: over to Marisa for details on current state

<wendyreid> https://w3c.github.io/sync-media-pub/

marisa: What had branched off pwg was a new WG
… This group looked at different sync approaches
… today we have SMIL in epub (sync media for narration)
… Lots of people dislike SMIL
… But at the time it was the best choice
… I no longer think it is the best choice
… We have a proposal for using WebVTT (see link in agenda)
… WebVTT is used for captions, but you can also use it for general sync

<ivan> The document Marisa is referring to

marisa: Outside of publishing no one uses SMIL
… webVTT is widely supported in browsers
… It might be reasonable to use in EPUB, though our group didn't look into packaging, etc
… Most compelling reason is that someday we should move forward. SMIL has roots in the 90s as a format for the digital talking book, but the world has moved on to other standards

George: SMIL is a W3C standard, but it comes from before the 2 implementation rules

ivan: Technical question - you refer to html 5, is webvtt dependent on html 5?

marisa: Without doing any research, I would say yes because it is on the media element

ivan: But media is also in xhtml

marisa: I don't know. Seems like it would be fine

ivan: In the charter there are some items that are not necessarily going to be recs in this charter, and there are 3 items in that group. The biggest is whether HTML will be allowed in epub (as opposed to xhtml)
… Should we add this work to the charter?

AvneeshSingh: Marisa has done a lot of work in the sync media group and it is always sad to me to see that there wasn't much traction
… As the COO of Daisy it would be helpful to know if there is a real business need? Will there be traction this time?
… It may be hard to get Reading Systems to switch
… Even Daisy was hard to convince to switch to MO over an older SMIL version
… Also, when we see sync media these days, we see it as useful in more places than epub?
… Sync is much broader, so maybe it doesn't belong in the epubwg

marisa: Asking people to switch is probably not going to be popular
… I think there is a case for having parallel formats
… existing implementations would still be fine

shiestyle: Can we just do this as a note?

marisa: This would touch on packaging, but I don't have details

wendyreid: We would have to add webvtt as a core media type
… Going back to the business need, it is hard to tell. MO books have always been around, but there isn't much growth
… one reason is difficulty in making the books. This may help that, as there is more tooling
… Also new challenges with TTS

ivan: On SMIL and webvtt co-existing, we don't have a choice. The charter requires that old features cannot be removed, so SMIL cannot be removed.
… Minor thing - from w3c perpsective, going from note to rec is now very hard.
… On packaging, do you mean we need to expand to include webvtt, or do you mean the structure itself has to change?

marisa: I would NEVER suggest we change the format of our most excellent packaging format
… Just some issues around what what needs to be added to the package document

George: I understand modernizing - do we get new features?

marisa: Not necessarily, but maybe
… One example is a sync of a poem. You can have word, sentence, and stanza sync
… Also when pointing at text, we have always used fragment identifiers which means they are everywhere in the content. wbvtt allows other ids.

marisa: On production side, it might be easier to use webvtt.
… I just saw a demo of using AI for this, which will probably be getting easier

ivan: In regards to finer selectors, this may synergize with annotations. New selectors are part of that chartered work

ivan: Do we agree to add this to the charter? Without obligation to produce a rec there isn't much of a barrier

marisa: First thing to look at is some caveats in SMIL -> WebVTT
… e.g. there are some SMIL docs that are technically allowed, but don't work in webvtt. That should be top of the agenda

wendyreid: At the very least we should add it to the charter so we can have further discussions
… No harm, but it at least signals we want to look at it
… But also maybe mention it to the BG

Hadrien: Looking at the draft proposal, there are some differences.
… SMIL is a tree with roles, which allows skipping and escaping. I don't see it i webvtt. Is that by design, or not possible, etc?

marisa: Section 4.3 covers skipping/escaping
… I think these features should be derived from the html
… Not really needed in SMIL
… You could do it in webvtt, but it would be better to get it from the text content

George: When DAISY added SMIL it was intended to replace audio books. Audio was the content (audio was king).
… The fact that we have text as the main content driver is better

Hadrien: I agree in some cases
… e.g. TTS
… but this is trickier since you have to go through two levels to get structural info

wendyreid: This is just a discussion for incubation, but these are all great topics for the chartered group

ivan: Thank you to the chair for getting us back on track
… Since this has to go to the AC and we want to avoid issues with that
… I think we should present this as bringing in webvtt and not making something new
… "Adding webvtt functionality to epub" or something like that would be the entry
… is that acceptable?

Hadrien: I am not convinced this is the right way to go. Narrowing to just webvtt is a missed opportunity. We should considered business arguments and features
… For example comics and a11y might be something we could address with a wider charter

ivan: I understand and intellectually agree. But complex and open ended charters will lead to "make a CG and incubate there" answers from the AC
… Here the message is taking an existing spec and combining it with epub 3, which is more managable for this group

wendyreid: Ivan is right we are looking at a specific tech, but it is also good to phrase this around sync media in general
… This is also about audio books, etc. So a broader phrasing might be better

AvneeshSingh: I was also thinking on broader lines
… and use this tech as an example instead of only focusing on it
… Then all sync media topics can be handled by this group

Hadrien: We already have an item in the charter that addresses my concerns. So I am connecting it back to things we already have in the charter instead of adding a new item

wendyreid: Sync is applicable to fixed and reflowable. Ivan and I will work on a PR for this in the charter

<wendyreid> readium/guided-navigation

wendyreid: Do we want to talk about the agenda comments from Hadrien?

hadrien: Our work is a little different from what Marisa did
… We looked at business cases and ended up at guided navigation. We can parse SMIL (or webvtt) and present them in an abstract way
… Thinking in that way has led to changes in other areas
… Not trying to realign with the web, but instead have found other real issues we can address with the same work
… e.g. accessible comics

wendyreid: The incubation item we add could cover this as well

ivan: The proposal Wendy made before should cover this
… We will put in a PR, please respond in the next few days so we can go to the AC withing a week
… AC vote will likely go to end of January
… So we would have a new charter in February unless things go sideways

wendyreid: Do we need a call for consensus to move ahead with the charter?

ivan: Not really needed

wendyreid: We will send out the charter, please comment now with issues, concerns, etc
… we want everything ready before we go to the AC
… we are moving next month's meetings out 1 week
… We may need to discuss timing and frequency