Meeting minutes
Multi-granularity highlighting in Media overlays
<Wendy Reid> w3c/
GeorgeK: looking at 3 names? (word, sentence, and paragraph), or something else? span around each word? for word level sync?
Wendy Reid: depends on how it is authored. we have an established pattern for wrapping and might be a word or sentence. now introducing nesting.
… we need to only define 2 bigger levels.
GeorgeK: paragraph then sentence inside the paragraph and word inside sentence. do you need id's on the words?
Hadrien Gardeur: first "sec" element equivalent of paragraph 2nd "sec" element we will call it "sentence" for now. and we don't need an epub type a the word level. you need id's for all of the levels.
<Hadrien Gardeur> Authoring Media Overlay with text fragments instead of an ID: w3c/
Wendy Reid: proposals are: chunk-large/small text-granularity-large/small
… internationally "chunk" is maybe ok.
Hadrien Gardeur: "text-chunk"
GeorgeK: what about "paragraph" and "word" group?
Hadrien Gardeur: we discussed those but we don't like this due to it might not be a paragraph or sentence.
… we want something more generic? but understandable.
Ivan Herman: adding 2 terms to a non-normative so we can change it if its wrong so no problem if we need to change it. I would just pick one of them and move on.
Charles LaPierre: I like "text-chunk" myself, it probably makes the most sense, followed by small/large
Hadrien Gardeur: this would be used "common language"
… concept of chunked encoding in http.
… few words or one word. Not a big issue to be used here for something similar.
Ivan Herman: this is something used by humans not only by machines. I tried to translate chunk into a few other languages. so I am not sure how to translate it to Hungarian thats why I like text-granularity myself.
Susan Neuhaus: text-chunking is a concept in how to teaching people to read, which makes it appropriate in a reading context but worried about translations.
Shinya Takami: Japanese is complicated language character base, word, sentence and paragraph so 4 degrees. so difficult. so i think small and large is fine.
Matt Garrish: tagging this what is it for words, sentence, paragraphs, can it be different things for small, or large, but wonder if it is a better approach than word, sentence, paragraph vs. a generic. I worry that it might not work in the end.
Dale Rogers: instructional/web design chunking is categorical. if a computer needs to sift through content, chunking doesn't resolve to a number. if a computer needs to resolve to a number.
Susan Neuhaus: can we trust Reading Systems that we author in and RS can localize it for their readers?
Hadrien Gardeur: RS will need to localize it to something their customers understand
… depends on how its structured, it will b a bit vague, we just need to agree on something.
Wendy Reid: by giving this a name, if I am a dev. and I want to group/chunk my content how should I do it / extent. To avoid phrase or clause, and paragraph is more universal. "section" could be a chapter, we don't want a MO for a large group of text like chapter, encourage authors to make them smaller so my proposal is "phrase" and "paragraph"
Hadrien Gardeur: best to make it more understandable or more vague. we just need to name it at this point.
Dale Rogers: "group" we use in industrial design maybe instead of "chunk"
Hadrien Gardeur: if you don't say anything about "text" it could be not understood. phrase, paragraph, sentence, but we need to make sure this is for text and there are sizes.
Ivan Herman: lets move ahead text-granularity small/large
… term we can all live with?
Wendy Reid: maybe offline?
Ivan Herman: didn't lead anywhere
Susan Neuhaus: granularity is what I have problems with.
… what about "range" instead
Hadrien Gardeur: I am fine with range
Ivan Herman: I am fine as well.
GeorgeK: text-group or text-range is good for me.
Ivan Herman: text-range small/large? works?
Wendy Reid: opposition?
Proposed: Name the text granularity options "text-range-small" or "text-range-large"
<Ivan Herman> +1
<Shinya Takami> +1
<Charles LaPierre> +1
<Gautier Chomel> +1
<Susan Neuhaus> +1
<Hadrien Gardeur> +1
<Toshiaki Koike> +1
<Wendy Reid> +1
<Dale Rogers> +1
<Masakazu Kitahara> +1
<Matt Garrish> 0
RESOLUTION: Name the text granularity options "text-range-small" or "text-range-large"
Matt Garrish: no paragraph then in this proposal? so is large paragraph? ok for now.
<img> in SMIL
https://
Hadrien Gardeur: Summary: provided examples we need something like this or based on this. move fwd with "At Risk"
… working on a PR
… need full examples
… created a CC comic 8 pages full script full audio we have panels and textual / audio descriptions speech bubbles etc.
… accessible comic at this point.
… epubCheck or in Spec. the more I dig into the spec the more I found issues in MO in general. images in spine I am not worried about the rest is what is concerning the content model.
… image in SMIL, at EDRLab iOS implementation for MO so I see this also to implement this in general. I am worried, about the current spec than the PR. ref. to the issues filed.
… PR 1 text element with 0,1 audio element.
… 2 combos text or text/audio now you can have text/audio, text/image, text/audio/image so 2 new additions.
… in Comic you can have descriptions of a panel or speech bubble.
… image/audio/text so you can have pre-recorded audio.
… braille tablet could use the text. or you can use TTS using the text.
… 2 issues from epubCheck image is not authorized which is expected. in my example would work best to keep script "text" eq. as a separate file and it shouts at me cause I am using MO on an image.
… one is expected the other less so.
… maybe without images in spine but that would be weird then script in same page and try to hide it. what I have done is more accessible than trying to hide things using css
… I opened them in multiple RS, thorium Apple books, very few that support MO anyways.
… Reading Systems I am not worried about that spec.
… now I have 4 examples, image text/audio, image with text equivalent. single script in a single file, and Media overlays in a spread.
… general impression we wrote the spec with FXL in mind but we forgot about spreads or content drawn across spreads, Reflowable with MO.
… matt you were worried about this in Reading System spec, I am not worried myself.
Matt Garrish: I haven't gotten a chance to look at this again. sure we can put an element, but what additional authoring req. say for split screen, not sure how it will work yet. roles needed where you have things in the spine etc.
… I don't have strong opinions on your other issues. Daniel and Marisa were the ones at DAISY taking charge on this work.
… very focused on XHTML document and need pros need to be reworked and implications with SVGs etc.
Hadrien Gardeur: timing issue unrelated to this PR. bigger issue with timing if you don't have audio what does it mean. introducing images doesn't affect timing issues.
… I haven't changed the content model for text.
… I don't think we can include duration and text and I don't think this PR makes it worse its already there.
… the spec is very open, doesn't limit fragment ID any fragment if you use any other fragment it is unclear.
… compatibility the current RS are broken, spreads are broken, desktop gets into a loop.
Matt Garrish: without TTS, time to show the text is problematic if there is no TTS. if we prerecord the TTS and have the RS do the TTS instead of the pre-recorded. We need some timing base. WHat happens when there is no audio. user needs to hit fwd to make the content move.
… images on there own can be problematic here.
Hadrien Gardeur: in the PR you can't just have images alone.
… timing I don't see the usefulness of the duration of each SMIL file as a RS will throw away. TTS on Text this PR makes a case for it.
… comic with textual equivalent can be displayed or sent to SR or Braille Tablet, using TTS with the text element this PR is a much stronger use case.
Wendy Reid: all good points.
… we need to reach out to Daniel and Marisa's take on this. as our MO Expert
… she has enhanced synchronized content. I will reach out to her.
Ivan Herman: My question / worry the problems you found in MO is it problems in SMIL original spec or how we took SML into EPUB?
… that spec is mainly used in EPUB. we don't own that SMIL spec. there is a problem there I see if there is a real problem with SML then we need to break the ties and implement independently.
Hadrien Gardeur: the latter, there only one case where issue there is content across the spread we may be limited by the SMIL syntax, all the other issues are in our spec.
… we may be able to tackle that another way, submitted an issue for that.
… there is a lot of interest in this space. this is part of our spec we haven't touched for a while. Serving the community I think doing additional work for that part of the spec will help everyone.
Susan Neuhaus: I know CSS is pushing into audio, can we use that in addition to SMIL?
Hadrien Gardeur: not really thats more for TTS and is a very complex discussion.
Wendy Reid: we need more review on this.
Hadrien Gardeur: last question, what should I do next?
… I can work on the PR to handle the remaining things or documenting things I am finding. open question.
… I wasn't expecting to open all these issues initially.
Wendy Reid: we will go through the issues. so document things as you run into them. we need review from Marisa on the PR / issues filed.
… and Daniel.
Ivan Herman: biyearly issues with Time change 3 week timezone changes.
… Europe this means 1 hour earlier. Japan 1 hour earlier as well.
… March 12 1 hour earlier, 2nd April goes back to normal.
… emails in the agenda will flag it.