APA + EPUB 3 Joint Meeting — Minutes

Date: 2021-10-28

See also the Agenda and the IRC Log


Present: Rick Johnson, Wendy Reid, Janina Sajka, Dave Cramer, Tzviya Siegman, Paul Grenier, Ben Schroeter, Sam Kanta, Charles LaPierre, Matthew Chan, Shinya Takami (高見真也), John Foliot, Brady Duga, Ruoxi Ran, Ivan Herman, George Kerscher, Bill Kasdorf, John Roque, Becky Gibson, Jason White, Victoria Lee, Avneesh Singh, Fredrik Fischer, Andrew Nevins, Jen Goulden, Deborah Kaplan


Guests: Brian Kardell, Victor Lopes, Shadi Abou-Zahra, Philippe le Hégaret

Chair: Janina Sajka

Scribe(s): Matthew Chan, Brady Duga


Janina Sajka: the topic is the request for horizontal review.
… we put 2 people on each spec to read.
… we didn’t find much beyond nits, some of which have already been filed on github.
… (not surprising because there is a lot of a11y expertise in epub wg).
… let’s give each of the spec reviewers a chance to speak now.
… i have some thoughts about the future of epub, and we can spend some time on that.
… we have some emerging specs that will be of interest to epub.
… e.g. support for people with cognitive and learning disabilities.
… (thanks to CharlesL for help on that, and to our several participants from the pronunciation TF).
… and will there be an epub 3.4? Or maybe some more major jumps in the future?.
… any more topics to be tabled?.

Avneesh Singh: we had some questions about respec.

Janina Sajka: i refer questions about respec to michael, maybe jason could speak to it.

Jason White: i’ve read the documentation and used it to an extent, not necessarily an expert.

Janina Sajka: we can talk about respec, but maybe we should try to get michael into a call.

Matt Garrish: we don’t have to solve respec issues today, but i’m curious what the process would be.
… who is responsible for changes to respec output? We don’t want to affect all w3c specs.

Janina Sajka: sure, let’s just discuss issues for today without needing to solve anything, necessarily.

Tzviya Siegman: there’s a group called spec prod, an email list. Try there. Michael would funnel questions to this group..
… one issue recently is about CR vs TR, and whether notes have a formal standing in W3C world.

Janina Sajka: and i think there are new vehicles: e.g. registries, W3C statements.
… notes could conceivably have more impact than they’ve had in the past.

Ivan Herman: the new process will become official on the 2nd Nov.

1. Reviewing the Specifications.

1.1. A11y document review.

Jason White: i filed some issues in github, but they are being handled.
… question about how this spec will adapt to future versions of WCAG a11y guidelines.

Janina Sajka: i’ve been working with SILVER, and i love the statement about a11y conformance applying to the whole book (not just a page or two).

Matt Garrish: difficulty is trying to anticipate what future of WCAG will be.
… we know it’ll be moving to a gold/silver etc. system.
… in the meantime we refer to the current WCAG standard.

Janina Sajka: speaking as APA chair, the biggest difference is that we’re going to try to build in a little wiggle room.
… so that the standard of conformance is not perfection, to allow for particular situations, to try to understand what role a spec plays vs policy decisions/legislation that happen at a government level.
… exactly how that will play out is still being debated.

Andrew Nevins: my issues in github are being tackled, thank you.
… one thing is that a script is being used to generate perma-links, i think mgarrish mentioned that this relates to a larger problem.

Matt Garrish: right, I hacked up a script that post-processes our document to solve the issue you alluded to, works for epub but its really something that should be fixed in respec.
… not sure that i’m the one that should be making those changes.
… so we have a temporary patch, but how much temporary patching do we want to do?.

Janina Sajka: it doesn’t sound like a WCAG issue to me.
… how often do you republish?.

Matt Garrish: every time there’s a PR being merged.

John Foliot: this is a google docs worksheet that is a working timeline for project SILVER.
… it suggests that they’re still 5 years out from completion.
… so i appreciate that we thinking ahead, but for epub we can probably stick to thinking about a shorter timeframe.

1.2. epub33 RS review.

Becky Gibson: didn’t find issues, but i wasn’t looking for respec type issues.

1.3. epub 33 core review.

Janina Sajka: i reviewed along with some others. No real issues, but I was surprised by some of the fixed format stuff.

Wendy Reid: FXL is a part of the industry that epub 3 made official. This content is primarily used for comics, print-replica content (e.g. cookbooks). We know about the a11y issues here..
… the WG is working on a note that gives guidance specifically on FXL a11y.
… the note is meant to add onto the core spec by suggesting the recommended way of doing things in FXL.

Wendy Reid: See Coming note on FXL.

Dave Cramer: with FXL its largely a way of taking advantage of things that already exist in HTML and CSS.
… and this stuff can be used well, or be misused from a11y point of view.
… that’s why we need guidelines.
… there’s also the added complication in epub of creating a spread (two different HTML pages side by side) that has no web analogue.

1.4. Review conclusions.

Avneesh Singh: can we officially say that epub 33 has passed APA horizontal review?.

Janina Sajka: i don’t hear any objections. There’s a checkbox that we need to check on our end, and that’ll be coming up soon.

2. Multilingual books.

Matt Garrish: and when there are books that are bilingual, sometimes there are books where content in both languages are displayed side-by-side.
… that’s a problem that isn’t unique to FXL.

Janina Sajka: yes, interlinear text can be helpful for language learning. I’ve seen examples of this in practice..
… maybe we can setup a working session together in the future.

Bill Kasdorf: The Harvard Univ. Press’s Loeb Classical Library is a really great example of English + Greek and English + Latin side-by-side..

Tzviya Siegman: janina mentioned a number of tools that do this sort of thing (e.g. Duolingo, bablefish).
… bible software is way more advanced in this field than general publishing.
… dictionary software doesn’t apply as directly.
… i know a handful of people who might have experience doing this sort of thing.

Janina Sajka: https://chaucer.fas.harvard.edu/pages/literary-works.

Bill Kasdorf: Harvard would have the resources and mission to help address this..

Janina Sajka: this is a link to a Harvard page that is an interlinear version of Canterbury Tales.

Charles LaPierre: this came up in a11y certification in Canada, e.g. a book that has both the english and french translation of a book.
… because publishers want to do this, we’re probably going to see this more and more.

Janina Sajka: i hope we can do this in a way that isn’t fixed layout.

Bill Kasdorf: In the Loeb the English is on the verso and the Greek and Latin is on the recto–facing pages aligning across..

George Kerscher: when we come up with a recommendation to the publisher on what we think works, maybe we could have some people review it and comment. Maybe the results could go into DAISY kb..

Tzviya Siegman: http://www.perseus.tufts.edu/hopper/opensource.

Janina Sajka: people have mentioned side-by-side, like in an opera libretto.
… the other way to do it is a line on top of a line.
… this offers more of a word-for-word comparison.

Ivan Herman: isn’t what you’re describing Ruby?.
… it was designed for far-eastern languages, but we might be able to use it for this as well.
… but i’m not sure if implementations of Ruby are accessible or not.

Brady Duga: there was an open letter to W3C from DAISY japan about Ruby, so I’m guessing it isn’t very accessible.

Wendy Reid: letter was just sent to us today, it’s about the inaccessibility of Ruby (link above).
… this is something we probably all want to work on, and especially if we are expanding use cases for it.

Bill Kasdorf: For the record I had that backwards–the Greek and Latin is on the verso and the English on the facing recto..

Tzviya Siegman: who was the letter sent to?.

Shinya Takami (高見真也): Makoto headed up the letter.
… Ruby is based on HTML and CSS.
… but maybe not a topic for epub 33.

3. The Future of EPUB

Wendy Reid: epub 33 will be the last revision that we intend to work on unless anyone makes a really convincing case for epub 34.
… we had a meeting yesterday called epub Next where we opened the floor to new feature requests, new business cases, etc..
… as we wrap epub 33 there will be a year of testing, but work on the spec will be quieter. So we can take this time to start incubating new ideas for the future of digital publishing.

Ivan Herman: any epub development is under a very strict constraint of backwards compatibility.
… an epub spec that breaks this backwards compatibility would cause serious industry backlash.
… we discussed many additional features for epub 3.3 - e.g. that content document must be strict XHTML.
… after much consideration we decided we could not change this now.

Janina Sajka: that makes sense, and kind of what I was imagining when I reviewed the fixed layout section.
… re. next gen: there was a suggestion that maybe instead of a building a bigger and better DOM, we iterate on what we have now and augment it for the use cases where it doesn’t currently work well.
… e.g. in braille representation of music.
… we might consult the wider community about nesting graphical structures (e.g. a way to drill down from a larger graphic to individual parts).

Jason White: it was broadly recognized at that meeting earlier this week that diagrams and statistical charts all present a11y issues.
… but we’re likely to encounter more of this in future digital media, once authors step away from print analog.
… in education, research, etc..

Fredrik Fischer: Nested graphics and objects - there is work going outside the w3c.
… part of working on braille, being done by American Printing House for the Blind (?).
… This is still confidential, but they would be interesting in talking about a sub-standard for braille, that might become a basis for digital braille.
… They are also working on nested objects which have a lot of issues, they would be interested on working on that with us.

Janina Sajka: I think I know who might be involved.

Fredrik Fischer: They have asked for this to flow through me, but I am a little out on thin ice.

Janina Sajka: I will look into this.

Avneesh Singh: DAISY reviewed the APH work and has already provided feedback.
… Everything mentioned about nesting, etc, it doesn’t look very epub specific.
… This seems more generally web related.
… But epub should participate in that.

Tzviya Siegman: Agreed this is larger than what epub would take on.
… But as we advance epub, we are making it more webby, so we will avoid anything that moves us away from the web.
… If we are working on nested images or changing the DOM, then that has to be a broader project.

George Kerscher: Agree with Avneesh.
… I have been asked to do a presentation on advanced features in digital books and immersive reading in VR/AR.
… No idea what to say, other than we want to follow the web here for things that are proven accessible.

Janina Sajka: Hearing that ee can think of good things to have, but we wouldn’t incubate and develop them in our group.

George Kerscher: We have heard that these documents are very important and need to retain the ability to open them in the future.

Janina Sajka: I think that was most of what I would want for epub 4.
… want to finish with a summary of what we are producing.
… Pronunciation is some immediate work that might be of value.

Wendy Reid: When we talk about the future of publications, they don’t need to be bleeding edge.
… we often hear that publishing is slow moving.
… epub 3 turned 10 this year, and we are still trying to convince people to adopt it.
… epub next is a long time horizon (years).
… Some ideas may not be new, they may just be improvements, or they may be filling wholes in what can be done today.

4. Review of work on pronunciation.

Ivan Herman: See Text-to-Speech EPUB WG Note.

Janina Sajka: Paul would you do a review?

Paul Grenier: Speech in HTML spec allows for aspects of SSML in HTML.
… We have a few proposals for bringing in SSML and socializing with AT and browser vendors.
… Still figuring out which implementation to go with, then will go to FPWD.

Janina Sajka: This is work to get consistent pronunciation across implementations. Just for the tough spots, don’t need it everywhere.

Paul Grenier: SSML brings a lot of options for full expression.
… We could do a full audio drama if we wanted with SSML.
… Trying to code the equivalent of a real world performer is tricky, but doable.
… Might be interesting for everyone.

Avneesh Singh: For this version of epub we are making sure we don’t get in the way of this work.
… in the longer term we are working on a note to cover this topic.
… Once this is mature, it will come by default to RSes.
… It shouldn’t require extra code for it.
… So we just need to make sure we don’t break it with our work on epub.
… Once it is in CR, all we have to do is point to the spec.

Tzviya Siegman: +1 to avneeshsingh.

Matt Garrish: I didn’t write that document! We wrote it, but no one came to the party.
… One of the reasons that note exists is it may be a blocker for epub 3.3.
… It isn’t competing, we just need to put it in a box somewhere.
… But when this work comes out we will definitely move to the new way.

Wendy Reid: One ideological shift is, when there wasn’t a tech available we just invented it.
… Without giving a thought to who would implement it. As part of W3C, we have to consider other groups.
… We can offer to help with our expertise, but should also rely on them for their expertise.

Matt Garrish: Part of the reason the note exists is for backward compatibility, due to our compatibility requirements.
… We aren’t trying to aggressively push it.

Janina Sajka: A11y often takes a while to become real.

Tzviya Siegman: One issue we have had with epub when we haven’t involved implementors along the way.
… The “when we spec it they will come” problem.
… Is there active involvement in the browser community for the APA work?.

Janina Sajka: Bad news - browsers are not in the task force.
… good news - when we had a joint talk with aria, it sounded like there were browser vendors who where interested.

George Kerscher: I attended a meeting last week with MS and the World Blind Union.
… Their read aloud tech would benefit tremendously from this work.
… read aloud is coming up more and more on discussions about neuro diversity.
… Read aloud is very important in these discussion, that would be a good reason to get this working in browsers.

Janina Sajka: Paul has pointed out that this can help PDAs do a better job.

Charles LaPierre: In that meeting it was mentioned there is a toolkit available for adding immersive reading.

Janina Sajka: Do we have any notes on that?.

George Kerscher: I will update after checking on it.

5. Personalization.

Charles LaPierre: Brief update.
… The spec for personalization allows content creators to add support for people with cognitize challenges.
… Personalization is farthest along.
… Can use it to add additional semantics.
… Goes further than aria roles.
… We have about 5 major areas in the module for actions.
… For example, clicking on buttons.
… There is also one for destination (where does this link take you?).
… Purpose: helps understand what certain fields are.
… and one for simplification.
… Symbols is one we might care about.
… Using symbols in place of words (?).

Janina Sajka: Turns out there are plenty of people who don’t deal well with a lot of text, but can handle it with some graphic signposts.
… Problem is meaning of graphics will depend on regional difference, etc.
… We are not adopting a universal symbol set.
… This is more important for interactive content, remains to be seen what it will mean for static content.
… And that’s all!.

Wendy Reid: Thanks everyone for feedback and reviews.

Jen Goulden: Wondering whether if we can arrange a quick chat with someone for help with participation.

everyone: Yes please!.