<scribe> scribe: Gottfried
Janina: Planned conversation on
publishing.
... Related conversation coming from RQTF about supporting
math, chemistry, and others
<scribe> New charter: https://www.w3.org/2018/08/apa-charter
John: Leonie encouraged other groups to join APA in a posting
Janina: That was a beautiful letter.
Michael: Everybody must rejoin
the APA wg because we have now normative work products.
... You will be kicked out if you do not rejoin.
... Everybody will get a prompt to rejoin.
Janina: Personalization tf is
moving from ARIA to APA.
... Work that ETS proposed on pronunciation for TTS.
John: All members of personalization TF will also need to join APA.
Janina: Started conversation with
Leonie about joints topics with web platform.
... One of them being personlization.
... Also architecture
... Publishing - need to dive into with Marisa
... Hopefuly we can outline the basic parameters on publishing
by Oct
... Also conversation on accessible knowledge domain
markup
... (still trying to find a good term for this)
John: Outstanding issues on media - programmatic access to transcripts, providing text alternatives to be poster attributes for video
Janina: Poster issue is an edge case. Other stuff has priority.
Gottfried: Talking about Mon-Tue agenda, or Thu-Fri?
Janina: Both. But will aim for
Mon-Tue as much as we can, to accommodate Gottfried.
... We can also swap with ARIA (Thu-Fri) if needed.
Janina: Wanted to make sure that
reordering is okay.
... We got lots of comments.
... Google's implementation is causing trouble.
... RQTF is already starting to address this.
... Our focus is not on privacy / security.
... Although we mention priv/sec in the doc.
Janina: We are delighted that
IDPF joined W3C.
... DAISY comes in as part of that.
... HTML5 allows to mark documents as digital publication.
Embedded audios and videos.
... Big opportunities, e.g. sign language translation (see
MAUR).
... A digital publication is not timed.
... But timed media may be provided as alternatives (e.g.
recorded audio)
... We should work our way to a position where we have various
types of media, each of which can function as the master
timeline.
... The master timeline may have alternative versions.
... The topic is publishing, not transcript.
... The transcript use case will be covered in the end.
John: Concern that reserved terms
are already being used and implemented.
... Printed document is media?
... Media has timeline and timestamps.
... We need more granular specificity.
Janina: Agree. We should stick to
what we agreed before for HTML5.
... They were confused about sign language translation.
... We don't own that term. But we might nudge publishing and
web-wg.
... We need to learn to articulate this without getting
stuck.
... It worked when we came up with a master resource - made
sense to them.
... But don't introduce it too soon.
Marisa: I am with the DAISY
consortium, involved in publishing wg in W3C.
... Originally ePub wg in IDPF.
... In EPUB3 we have media overlays.
... The underlying format was SMIL - we defined a subset of
SMIL for media overlays.
... Goal is to continue this in the W3C publishing wg.
... But SMIL is heavy and complicated.
Janina: HTML5 killed it.
Marisa: We are looking at
existing technologies to replace SMIL for media overlays.
... Problem is the constituents of the publishing wg. We do not
have people that want to produce synchronized audio and
video.
... If there is demand of synchronized media in ebooks, let me
know.
... Audio description working group looking at TTML2.
<Zakim> JF, you wanted to ask about popcorn.js
John: Are you familiar with
Popcorn-JS (JS library)?
... PopcornJS looks at media asset, takes a fixed timeline,
fires events at runtime based on timeline markers.
... Genome project: proof of concept experiment
... You want to have a similar function but in another
format.
... Fire and access supplemental content at a specific point in
time.
Marisa: We are looking for
synchronized text in audio books.
... Technologically this is very possible in a browser context.
Use JS today.
... But format for the timeline is not clear.
John: External resource rather than inline markup?
Marisa: Yes
Janina: I think we need to keep
talking. Discuss potentially useful approaches.
... John and i have experience about what works and not in the
HTML5 world
... You can get all what you want, but make it more useful for
mainstream publishing.
... This merger will prove its value in a big way.
... DAISY book concept - synchronize audio and text - looking
into using HTML5 for this?
Marisa: We want to retire
SMIL.
... Either just use JS or TTML2 or so.
... Don't want to make it complicated for implementers.
... We do not need a complicated timing solution.
Janina: Your use case also
satisfies the transcript case.
... But talking about it now complicates things.
Marisa: We do not have representation in our wg for this.
Janina: If we have an audio
resource with transcript - that's a text version of the audio -
no need for synchronization.
... Not the most useful kind of transcript, but it covers
it.
Marisa: We are looking for finer-grained synchronization.
Janina: The most stupid use case can be a starting point.
John: Primary asset - other
things are related to it.
... Audio feed of text.
... Timestamped audio feed - highlighting possible.
... It is all about identifying different kinds of relations to
a primary asset.
Janina: I think only one can be a primary.
John: We need more to say about the primary asset - overlap with personalization.
Janina: audio descriptions,
captions, sign language, multiple languages, and so on.
... And here is the transcript.
... The alternative for transcript would use Marisa's markup
when there is synchronized audio.
John: We need a mechanism to say: Here is a primary asset.
Janina: Apple will push back.
Marisa: I understand what you are
proposing.
... But i don't know if it works.
... I will get back to the publishing wg, and they will want to
hear specifics about is.
John: Agree. We need a concept,
but also examples of implementation.
... Need to be clear on the larger picture of the asset, then
identify use cases.
Janina: Enough for now. Need to
continue this next week.
... We need to define this well enough to have a meaningful
conversation at TPAC with others.
... Clash with audio description group. We don't call it audio
description in the MAUR.
... I will jump in here, giving the people an overview on
MAUR.
... Naming clash, we disagree with ISO on this. Video
description or audio description?
... Comparison of TTML and SMIL?
Marisa: Looked at proposed AD by
Nigel and ??
... Not similar to how DAISY has been using SMIL.
... Timing structure is more intricate than what we
require.
... Completely different text format.
Janina: You may have to live with that.
Marisa: Publishing wg centered
around html5 format for ebooks.
... Maybe there are other serializations - discussed with
Nigel.
Janina: We need TTML or
WebVT.
... Text version needs to be HTML markup.
... If TTML is too fine - not a problem - define a subset.
John: First support of TTML in Flash players was only a subset too.
Michael: SVG is up.
... One of the timelines may be up, but i don't remember.
Janina: We will put action item review as a main point on the next agenda.
Page visibility
Janina: I will ask them for an extension.
Michael: If it is only new draft, it might be possible. If it goes to CR, might be a problem.
Janina: We need some developers.
This is scribe.perl Revision: 1.152 of Date: 2017/02/06 11:04:15 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Succeeded: s/markes/markers/ Succeeded: s/thigs/things/ Default Present: janina, marisa, gottfried, MichaelC, JF, IanPouncey Present: janina marisa gottfried MichaelC JF IanPouncey Regrets: Leonie Found Scribe: Gottfried Inferring ScribeNick: gottfried Found Date: 08 Aug 2018 People with action items: WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]