Meeting minutes
<wendyreid> date: 2025-02-27
WASM resolution (w3c/epub-specs#2686)
Wendy: We have a packed agenda today
<wendyreid> w3c/
Wendy: We would like to have a resolution affirming 2686, which includes webassembly…
<wendyreid> Proposed: Merge PR 2686.
<Sueneu> +1
<wendyreid> +1
<gpellegrino> 0
duga: There were some changes from out discussion…
something of a backdoor for fallbacks for embedded content. We’ve closed an open door and removed embedded content
mgarrish: mostly it maps to what we talked about, added some clarifications and examples…
<CharlesL> +1
there is a summary in the PR
<shiestyle> +1
<duga> +1
<mgarrish> +1
<toshiakikoike> +1
<MasakazuKitahara> +1
<George> +1
RESOLUTION: Merge PR 2686.
Update from FXL Accessibility
Wendy: Next item update from FXL a11y TF
… FXL TF has been working on a recommendations document for creating FXL accessible content and recognizing the limitations
… we’ve been floating some ideas about how we can address the limitations of FXL accessibility
… Hadrien has mentioned a reading system that can display content in different ways
… including possibly rendering the FXL content as reflowable
Hadrien: What we have in mind is experimental. We are trying one platform first…
… for this we are doing it on the web, because of our opensource web reader, Thorium web
… we would walk the dom to extract structure and utterances and pass them along. Here we would use that differently
… since in FXL words and letters are often separated, we are thinking of putting the text together
… and creating a readability mode that allows font changes, colors, etc
… this will depend on the quality of the content, if it is well structured, you will get something readable
… on larger screens we will do this side by side, you will see the fixed layout on one side and the reflowable on the other
… on smaller screens it will be a toggle
… hard to say when we will have something to show, possibly in the summer or fall
duga: how does this relate to the work the taskforce is doing?
wendyreid: since we’ve identified things keeping fxl layouts from being fully accessible this is one of the ways we are looking to solve this
… on the practical side, we are writing a techniques doc, about how to achieve the recommendations
… the goal is to encourage ebook creators to write the code in a way that allows a mode like this to work
duga: Has there been any exploration of using LLMs to extract useful data?
wendyreid: we have not
George: could a fxl book be validated to be useful for a reading mode?
wendyreid: we could run the book with ace and get some of the results, but it would still require human review for things like reading order
George: does the technique thread the content through the page?
wendyreid: yes. If they follow the recommendations.
Hadrien: Assessing if content will work is possible
… we know which languages are used from the metadata, we can extract the text
… we can check and see if the results don’t look like the indicated language, we will know something like the reading order is wrong
… about LLMs its interesting to see what is happening in screen readers
… they will pick up alt text, and if that is not available it will create a description
… I’m not sure all the APIs we need are exposed, we can’t always use them
… but it is worth exploring. It may or may not be better than walking the DOM
… lots of universities are focused on this area, uses like extracting texts from comics
… LLMs are working better than before
… the challenge is making a generalized model and using the embedded software
CharlesL: Most fxl books can reach WCAG A, but not AA
… Benetech is working with Amazon’s publishing to certify their pipelines, looking for ways their fxl pipeline can reach WCAG A
What to update in EPUB A11y 1.1?
Wendy: We started this discussion but there is still and open question
… what do we need to update in accessibility? are there areas we might look at?
AvneeshSingh: We discussed accessMode sufficient metadata, there is a demand from the APA to stick with 2.0 for the floor
… but we should move forward from that spec
… alignment with 3.4 we don’t know what that would be but are keeping our eyes open
… annotations, webtoons, we need to inquire what is required for accessibility for these things
Dark mode guidance for reading systems (w3c/pm-wg#14)
<wendyreid> w3c/
Wendy: Brady brought up the issue of day/night modes in reading systems
… this has been an issue for a long time, more and more publishers are taking a look at this
duga: I had a very specific issue, color space meta information in documents, that can change the appearance based on reading mode
… I’m curious what happens when we add this to an EPUB and how it works in the reading systems
… I don’t want to recreate a new system for day/night mode
… but we might need to advise people on color mode in their documents
Hadrien: I don’t think it is a good idea to use this when your authoring content
… we have apps were the browser or system settings will apply a different theme for the app but not the content
… some apps might default to the system setting, but not all of them, not all users will have a preference
… content creators should steer clear of this
<duga> +1
Shiestyle: in Japan we use characters that are image based. We have to prepare both light and dark mode content
… that makes Japanese particularly challenging
duga: you could do it with a filter in your CSS, but Hadrien is right, this could cause mapping problems
… the next step would be to generate some content and see how reading systems respond
… I wil do that
Adding dark mode styles to EPUB 3.4 documents (w3c/epub-specs#2685)
<wendyreid> w3c/
Wendy: we have the APR open to make the spec visible in dark mode
<CharlesL> +1 to merge
Wendy: can we approve dark mode APR to be merged?
<wendyreid> PROPOSED: Merge PR 2685, add dark mode to EPUB 3.4 drafts.
<Sueneu> +1
<duga> +1
<tzviya> +1
<wendyreid> +1
<shiestyle> +1
<romain> +1
<toshiakikoike> +1
<MasakazuKitahara> +1
RESOLUTION: Merge PR 2685, add dark mode to EPUB 3.4 drafts.
Open letter about interoperable ebook standards
<wendyreid> https://
Wendy: Did anyone see the open letter to interoperable ebook standards?
… this came from academia
… they want to look into ease of use, ease of discovery, and accessiblity
… this is an opportunity to engage with them
Tzviya: matt, would some of your connections from the library world have connections there?
Hadrien: they are asking for two things I
… I’ve created. They are familiar with EPUB
… their problem is having to pay more and more for content and reading systems
… they would like to be able to create one collection, to harvest content into one collection
… this is about fragmentation of discovery, and fragmentation of the reading experience
… this has already happened in a few countries
Shiestyle: This seems not from academia but from libraries
… this article says Readium only, I don’t understand the objectives
wendyreid: I’ve experienced this with my own studies. From my own university library, I could click through to 20 other sites and may or may not have access to the material
duga: Is this article looking for a rights management system?
Hadrien: they aren’t looking at rights management, but want a single collection that can be used by whatever system you are using
… you could imagine a Kobo reader that supports DRM being able to read across ebook vendors
… its about giving end users options
<Zakim> tzviya, you wanted to identity issue
Hadrien: using standard interoperability would let the end user choose their reader
tzviya: this is also about user identity management
… the rights may not carry over from one library to a connected one
… this is complex, but there are some cobbled together solutions
Hadrien: There is an identity issue moving from one platform to another, but if you only need to go to your university/institutions platform, it would be easier to solve
… identity issues could be largely solved by bringing everything to one platform
shiestyle: we created a digital comics task force, let me or Hadrien know if you are interested