W3C

– DRAFT –
FXL Accessibility Meeting 11 July 2023

11 July 2023

Attendees

Present
CharlesL, CircularKen, duga_, gpellegrino, jamesY, jgriggs, Naomi, sue-neu
Regrets
-
Chair
wendyreid
Scribe
duga_, wendyreid

Meeting minutes

wendyreid: Compaints about Toronta public transit

<wendyreid> https://github.com/w3c/publishingcg/wiki/Fixed-Layout-Accessibility-Taskforce

wendyreid: Summarized matrix of features/formats - see link above
… the original document distilled everything to 4 basic types that appear in fixed layout
… There was way too much in illustrative works (comics, illustrated books, etc)
… At least need to break out comics
… And now we have an empty column for proposed solutions
… because different content needs different solutions (e.g. comics vs text books)

sue-neu: Agreed, especially with comics where you may want to rearrange panels

wendyreid: There is still some more breakdown we can do
… So what are the solutions?

gpellegrino: For comics I don't really know, maybe Hadrien does. But for the first one I would say HTML or SVG techniques
… Same for the second one
… and maybe also the third one

sue-neu: For children's picture books, something that came up was hybrid epubs
… where you can convert from fixed to reflow
… Also multiple renditions

Naomi: There is also mixed fixed/flowing
… where only the pages that need to be are fixed
… both this strategy and multiple renditions have been around forever
… But if you COULD make it reflow, why not just do only that?

gpellegrino: I think it is possible to mix fixed/reflow
… For the first 3 I would add we could add some algorithm to generate reflow for that (automatically done by reading system)

wendyreid: Visual/textual has been proposed multiple times
… Maybe instead of a book feature, maybe make it a reading system feature

Naomi: I still want control over how that content looks
… Need to be clear what formatting has to stay (or has to go)

sue-neu: What is visual/textual idea?
… Where are the docs?!

<wendyreid> w3c/publishingcg#60

<wendyreid> w3c/publishingcg#23

sue-neu: If you can make fixed reflow, why not? Because for early readers you may want fixed even if it can be made reflow

Naomi: Agreed for kids books, but cookbooks, etc it seems odd

Naomi: This is a good case for fixed and flowing

duga_: Wanted to generalize what Naomi was saying about mixed fixed and flowing
… we did it with multiple renditions and it didn't go anywhere
… we discourage FXL, but if we gave tools to make that possible, and if we got RX implementations of mixed fixed and flowing, that would show publishers
… it's an important path to go down, we need to give content creators tools to use FXL sparingly

gpellegrino: Guidance on fixed to reflow - a well known fixed layout format has provided a way to create html from their format
… We should be able to do that

CharlesL: This idea of visual vs textual view, this is something that was already done ages ago
… Pushing back on styling, for someone who needs the text they often want to drop the styles
… so textual needs to be truly just the text so it can be styled as I need it

gautier: Working with specialized readers, there are some complexities to add formats
… Something that is working is fallbacks
… for instance on devices with no screens
… This isn't an edge case, we should take into consideration this working solution (fallback to xhtml)

CircularKen: I like fixed layout! It can be very useful for well designed pages to exist
… Maybe a simple CSS switch could be used to leave the basic styling of the page (bold, italic), but drop the fixed layout stuff (positioning)

<CharlesL> +1 to Basic CSS (could work :)

wendyreid: I don't think anyone wants to get rid of fixed layout

CircularKen: Don't discard the meaning of remaining in the fixed layout page

wendyreid: A11y doesn't just mean visual - layout can be important for some a11y use cases

sue-neu: What do you mean by CSS switch? media queries or amp?

CircularKen: Not an expert, but some sort of switch when you open or read the book
… Having a single version you can just switch between seems very elegant

wendyreid: I see a switch like this as a reading system feature

duga_:

duga_: CSS switch has been done a few times with varying degress of success
… day/night mode, vertical/horizontal mode
… alternate stylesheets
… it was horrible and hard to implement
… we implemented horizontal/vertical, but it was intended to be user-driven
… the other place more successfully was media-overlays
… we inject classes into the content to cause the content to be rendered differently
… the reading system renders the book based on the CSS provided by the content creator
… simply by adding a class to the root element

duga_: I don't know about implementation, I think it was a bad feature
… I wouldn't recommend anyone implement today

Naomi: Could we use the same mechanisms to do a reflow/fixed switch?

duga_: No, the way we did it was a matrix
… completely unrelated things, but needed all these stylesheets
… but it doesn't make much sense
… if we added another feature, the matrix gets even more complex
… maybe media-overlays, much better support

CircularKen: I know one likes visual to textual, but it is better to fixed layout to reflowable
… It is really prepaginated to not prepaginated

wendyreid: Prepaginated and reflow are global but they can be overridden for each element
… at least that is how it is written now
… Maybe this is global?

duga_: I don't think it has to, always view pre-pag as a global default, every individual item as a toggle, but the default is the global setting
… just an impression
… I don't know if this would need to be a global property, I don't think the content would suddenly would not be pre-paginated
… if i had a pre-paginated page, and I changed the CSS, it doesn't get merged into the surrounding pages
… it feels targeted, this page has this CSS override
… every page could have it's own property for that
… when you enter the reflow mode, every page gets the CSS for reflow

sue-neu: I agree this could be a local setting vs global
… maybe comparing pages across media types (e.g. knowing the page you are on in a classroom)

wendyreid: Yeah, wouldn't want to merge pages
… What meant about global vs spine, the default is reflow
… So you have to specify pre-paginated

The sooner you can tell the RS what the features are the better

duga_: I kind of agree because I did it that way, I also disagree because it backed me into a corner with not supporting a feature I wanted to, mixed fxl and reflow
… now I assume everything is that global default
… I disagree that it's any harder by spine item, you better use the spine to know what is coming
… RS reading the spine
… I would agree if it was buried in the content itself, since it's late in the game
… if it's in the manifest

wendyreid: I would be happy with a global property that says it is mixed

duga_: Not sure if its necessary, just walk through the spine

CharlesL: Yeah, I guess. Knowing if it is one, the other, or both, would be helpful for libraries and bookstores etc

duga_: I believe that metadata is already there and can be used by libraries/bookstores, but if they are not, it's their problem
… you can parse the manifest to find the values

CircularKen: What is the easiest thing to do? Is book level the easiest? maybe do that first?

wendyreid: From a spec perspective, the easiest is to say that it already exists
… Job done!
… Currently the spec isn't terribly clear that you can do mixed, though it is allowed
… Adding a new global property is hard. We are not chartered to do that
… So using existing things will be easiest

duga_: Wanted to point out the danger of the easiest path, if we were to go that route, we would need to discard per-item reflow/fxl switch, we'd need a global that meant "this entire book is pre-pag/this book is reflow", it mightthe right thing to do from a reality standpoint
… it's not a case of the easy one first, hard later, it's a case of doing the ???

duga: Going down the easiest path means we will likely lose the ability to have mixed content. So easy might be instead of hard, instead of the first step to it

CharlesL: I am hearing we can't rely on metadata, we should be walking the spine
… If we need to walk the spine, maybe we can get DAISY to update ACE

duga_: I don't think it's really supported yet, it might be premature to address the metadata
… if it becomes a thing, that makes sense.

wendyreid: MORE COFFEE!

Naomi: Does epub test still have mixed fixed/reflow content?

wendyreid: Not sure, but there is an epub 3 test file
… They seem to pass, but we are not sure exactly what is being tested

w3c/epub-tests

https://github.com/w3c/epub-tests/tree/main/tests

Naomi: We should test this, so far my adhoc tests show it isn't supported

fxl-spine-overrides

wendyreid: Over time
… see you in 2 weeks

Minutes manually created (not a transcript), formatted by scribe.perl version 210 (Wed Jan 11 19:21:32 2023 UTC).

Diagnostics

Succeeded: s/Ken/CircularKen/

No scribenick or scribe found. Guessed: duga_

Maybe present: duga, gautier, wendyreid

All speakers: CharlesL, CircularKen, duga, duga_, gautier, gpellegrino, Naomi, sue-neu, wendyreid

Active on IRC: CharlesL, CircularKen, duga_, gautier, gpellegrino, jamesY, jgriggs, Naomi, sue-neu, wendyreid