Meeting minutes
Resolutions from last week
Finalize resolutions from last week
Wendy Reid: following up last week discussions, about going to ISO.
<Wendy Reid> PROPOSED: Once the liaison with SC34 is established, we will submit EPUB 3.4, EPUB Reading Systems 3.4, and EPUB Accessibility 1.2 as drafts for their review.
<Murata Makoto> +1
<Charles LaPierre> +1
<Murata Makoto> Hopefully in May.
Wendy Reid: it's about sending a draft to speed up later process.
<Dale Rogers> +1
RESOLUTION: Once the liaison with SC34 is established, we will submit EPUB 3.4, EPUB Reading Systems 3.4, and EPUB Accessibility 1.2 as drafts for their review.
<Toshiaki Koike> +1
<Avneesh Singh> +1
<Gregorio Pellegrino> 0
<Wendy Reid> PROPOSED: We will submit EPUB 3.4, EPUB Reading Systems 3.4, and EPUB Accessibility 1.2 to ISO as part of the PAS process when all documents reach recommendation status.
<Murata Makoto> +1
<Charles LaPierre> +1
<Masakazu Kitahara> +1
<Romain Deltour> +1
<Dale Rogers> +1
<Toshiaki Koike> +1
<Avneesh Singh> +1
Wendy Reid: this is to agree on submiting the final W3C reviewed once it is done.
RESOLUTION: We will submit EPUB 3.4, EPUB Reading Systems 3.4, and EPUB Accessibility 1.2 to ISO as part of the PAS process when all documents reach recommendation status.
Rechartering
<Wendy Reid> https://
Wendy Reid: those are a draft and a diff.
… we reworked the scope. and the timeline We may want to include other things.
Avneesh Singh: the timeline looks tight, we may need more time. Also I see conflict between adopting new technologies for sync media like WebVTT and maintaining compatibility.
Ivan Herman: we changed the iso paragraphe, otherwise the rest is identical.
… It is not in contradiction because we are limited in what we can do. Removing completly SMIL is not possible, we are bound to backward compatibility.
Ivan Herman: anotations aside, my impression is that this is ready to go, some editorial changes but nothing significant. 3,4 can go to CR before summer hope we don't get blocking reviews.
Ivan Herman: regarding 3.5, we want to avoid maintaining two documents in parrallel, that's why 3,5 first draft should began after 3,4 is done. It is not a problem to start discussing 3.5 before.
Wendy Reid: 11 months between snapshot and rec is a lot.
Hadrien Gardeur: I'm not worried about the timeline for 3.5, some things have been discussed and are already implemented, tools allow to generate time fragments.
… I think a mid term face to face meeting would be good in the process.
… on annotations, we can discuss further if it belongs to 3.4 or 3.5.
Hadrien Gardeur: exploring text fragment is not tied to VTT in my mind, so I see no contradiction here.
George Kerscher: after discussion with ISO folks, it seems we should push to REC, Annotation should not slow the process
Ivan Herman: Annotation can bve separated from 3.4 or 3.5, we don't have to decide to tie it to one or another.
George Kerscher: I want to be sure we align the CG work to allow time for proper community feedback.
Ivan Herman: to be very formal, separating both would block the ongoing work on media overlay, that's why we propose this combined charter instead of waiting the end of this charter to create a new one.
Ivan Herman: I'm excited to hear about sign language, not sure how far we are from having it.
Hadrien Gardeur: the ongoing work is important to be done now, there are business reasons, people will implement, they need a standard or they will go there way.
<Susan Neuhaus> s/theyre/there/
Avneesh Singh: Daniel did this implementation in Thorium, we had identified issues, that's why we want to discuss it further, we are not 100% satisfied with the existing and want to do it the right way.
Avneesh Singh: I think it is a separate bullet with liaison with sync media CG. That would be a good way. Marisa has thing to say on this.
Hadrien Gardeur: I see 3 categories of items: fixing the spec; content model (actual restrictions we put on SMIL); and experimental things like text frament and video
… text fragment is solving cases with different layouts, people ar implementing that, so we'll have feedback soon. Video and image is different.
… probably thing to be shown at the end of the year because EAA put pressure on visual narratives.and I am quite excited to think about implementing video syncronisation.
Ivan Herman: Hadrien Gardeur, it would be ideal if you can make a PR on the charter with all that.
Hadrien Gardeur: works for me.
Ivan Herman: we all need to remember to not be too detailed in the charter, keep things open.
Ivan Herman: can we have a resolution to state we start with this draft?
Gregorio Pellegrino: I need a bit more time to review and comment.
Wendy Reid: I'll also send to the CG and BG to get feedback.
Merging Annotation Sets
<Wendy Reid> w3c/
<Wendy Reid> w3c/
Wendy Reid: not to be resolved today, we wait for Laurent, but we can discuss a bit. How can we be sure that annotations are for this book.
Hadrien Gardeur: it's probable that most RS will manage notes in local databases, and export on user demand. It's the reasonable way to have it implemented shortly, so anotations will not live aside. For book identification, it's complex, nothing magical, unique identifiers are not sufficiently unique. The RS may alert, reject or try something.
Ivan Herman: that's clear, how RS manage annotations is not our business. But we may need to explicitly state that in the note. Same for the behavior on correctness of annotation relation to a book.
Wendy Reid: still we should provide as much pointers as possible on how to handle those parts.
Susan Neuhaus: should we provide guidance on allowing users to delete an import
Hadrien Gardeur: there are several use cases, maybe I don't have book context, I import annotations from inside a bookshelf or from inside a book
… RS compare metadata, if it does not find, it can allow user to select a book. There need to be a second check to see if it fits, but in some cases, I may not have access to the book selectors, like if it is expired. Si software may have access to the file or not. Warning user is always a good idea. RS consider that the user is always right, he may have a good reason to do that.
… trying to do too well can lead to user complains
Ivan Herman: annotation set has it's own metadata, those should be kept, so this is important to realise.
Dale Rogers: will we have a validator for annotation package? Is it part of our process or responsibility of other groups?
Ivan Herman: we have JSON schema, that's all.
Hadrien Gardeur: if the software has access to files, it will not make much use of annotations metadatas, while for solutions that have no access to file those metadata will be important.