Must address Fixed Layout accessibility

[GeorgeKerscher]

The lack of accessibility in Fixed Layout titles must be addressed in the new charter of the working group. Whith Title 2 of the ADA going into effect in April 2026 and the increased demands for accessibility around the world, we must deliver to the publishing industry a solution.

While it is possible today to make a Fixed Layout title as accessible as possible, it does not make it good enough to pass any kind of accessibility test. It is also possible to make a separate title and put it into the market, but this violates one of the long-standing principles in the W3C that a separate version is never the appropriate way to deal with accessibility. Alternative versions are never maintained and a separate version is never as well done as the primary version.

For these reasons, we need to solve the problem. Multiple renditions was thought to provide a solution, but that approach was not implemented. The WG has options for addressing this problem. It could be the resurrection of the multiple rendition approach or a new approach that the WG invents.

The resultmust be a solution in the package that contains an alternative presentation that will meet the accessibility litmus test .

[iherman]

@GeorgeKerscher, I am hesitant to add this to the bullet list of explicit deliverables in the scope section, because I do not think we have a clear idea what exactly should be standardized and how. On the other hand, there is a (smaller) list of items that puts certain topics in scope for further incubation without taking the commitment of delivering a recommendation in this round. At this moment, there is a bullet item on the A11y issues of Mangas/Webtoons; an idea would be to add FL either as part of that bullet item or as a separate one. WDYT?

cc @wareid

[HadrienGardeur]

For these reasons, we need to solve the problem. Multiple renditions was thought to provide a solution, but that approach was not implemented. The WG has options for addressing this problem. It could be the resurrection of the multiple rendition approach or a new approach that the WG invents.

I feel that multiple renditions is a heavy handed approach to that issue, and would come with its own set of problems (starting with the fact that multiple renditions has nearly zero adoption).

It's hard to properly categorize FXL publications, but let's try anyway:

• first of all there are FXL publications that should be reflowable instead, the layout might be simplified but you wouldn't lose much
• then you have FXL publications where image descriptions and text in the right reading order are mostly good enough, but that are lacking the ability to customize text display
• finally you have publications such as manga/comics/webtoons, where you'd need a much more granular level of description of what's in each image

For the first group, publishers should be creating reflowable publications instead, it would be completely overkill to use multiple renditions.

For the second group, reading systems could extract text and display it in a component alongside the text where we could customize text (that's what we've been calling a "reader mode" in the past).

For the third group, I believe that a fragment based approach would work best. I don't think that EPUB (or Daisy for that matter) truly offers a solution for this type of content. Multiple renditions wouldn't solve that issue either.

Here's what I've been working on for that last group: https://readium.org/guided-navigation/examples/Pepper%20%26%20Carrot/

[iherman]

@HadrienGardeur @GeorgeKerscher let us remember that this is an issue on the charter proposal.

The comment of @HadrienGardeur indicates to me that this problem cannot be the subject of the explicit list of deliverables and it still needs some incubation. Hence my proposal in https://github.com/w3c/publ-maintenance-wg-charter/issues/34#issuecomment-2410283211: as far as the charter goes, this should be added to the WG incubation entries but not to those that would be scheduled as a standard in EPUB 3.4...

[GeorgeKerscher]

If we can incubate it starting now, it should be able to fold it into 3.4, IMO.

[HadrienGardeur]

I can present my work in a future call if that helps.

I'll miss the November meeting of the working group (I'll be traveling in Japan AND November 1 is a holiday in France), is it too late if we discuss it during our December 6 meeting?

[mattgarrish]

Correct me if I'm wrong, but having it in the list of things we'll incubate doesn't mean we are forbidden from moving a proposal forward to REC track. It just means that at this point there's still work to figure out what we're even proposing.

That sounds like the current state of FXL accessibility.

[iherman]

Correct me if I'm wrong, but having it in the list of things we'll incubate doesn't mean we are forbidden from moving a proposal forward to REC track. It just means that at this point there's still work to figure out what we're even proposing.

That sounds like the current state of FXL accessibility.

Exactly, that is the intention. (That being said, the AC still has to accept this slightly "open-ended" approach.)

I still do not believe that we should commit ourselves to more than that at this point.

[wareid]

I would agree with having this in the items to incubate and explore, I don't want to promise there will be something in EPUB 3.4 addressing FXL accessibility that solves all problems relating to WCAG.