EPUB 3 Working Group FXL A11y Task Force Telco — Minutes

Date: 2021-04-06

See also the Agenda and the IRC Log


Present: George Kerscher, Wendy Reid, Ken Jones, Gregorio Pellegrino, Laura Brady, Matthew Chan, Charles LaPierre, Rachel Osolen, Brady Duga, Will Awad



Chair: Wendy Reid

Scribe(s): Matthew Chan


Wendy Reid: hi everyone, let’s start with updates from what we did last week

Wendy Reid: https://w3c.github.io/epub-specs/epub33/fxl-a11y/

Wendy Reid: biggest update is that we now have a document in github
… basically a skeleton version of the note, that we can now edit on there
… please make your contributions via github PRs
… or you can submit the text via email and someone will make the PR for you
… its in a very early state right now
… if sections are empty, then they don’t appear

Wendy Reid: https://kb.daisy.org/publishing/docs/contribute/

Wendy Reid: re. updates to Daisy KB
… mgarrish refers us to that link
… he is not very strict about it, so go ahead and open issues or file PRs

Gregorio Pellegrino: i tried to figure out what were the coolest new CSS and HTML features for accessible FXL epubs
… CSS flexbox, CSS grid, CSS media-queries
… trying to create some epubs that mix FXL and reflow, which is enabled via the specs, but not widely supported by RS or implemented
… using ARIA attributes to pull reflowable things from an FXL page

Ken Jones: talking with Lars from Colibrio, and he mentioned that the synced media group were working on doing panning and zooming within a page, like for comic books
… their project is tied into audio, but it might also be useful for us

George Kerscher: the epubtest.org site now has its 90 day work plan up, including an item for an experimental epub
… so we can start doing some AT testing with that

Charles LaPierre: the epubtest.org test books are in a very specific format
… i’d be happy to take the experimental book that we create here and create the test-version of the book
… and write the test cases

Ken Jones: at the moment if an FXL epub is given to an AT screen reader, what happens?

Gregorio Pellegrino: i didn’t test it, but usually AT intercepts HTML
… so what the user gets is the actual HTML content
… another set of AT are magnifiers, in which case you get a zoomed view

Charles LaPierre: we work with a children’s reading project who are retrofitting children’s books
… books are basically image with a text box
… we basically just added alt-text to the page

George Kerscher: i’ve had some of the titles read out text, but it wasn’t clear whether it was alt-text being read or text in the content
… if the content text is in the correct reading order that should be pretty easy to read
… text in images is a different (worse) matter, of course

Gregorio Pellegrino: here in ITA, we’ve also asked publishers for sample content
… our big publisher may give us some content, but he’s asked us what our policies are
… something that he can show authors etc.
… about how their content will be used, where it will be used
… he also asked if we will use some kind of DRM, but I already told him “no”

Wendy Reid: okay, I’ll look into that
… I asked business group last week if publishers were interested in sharing content
… Lisa (PRH) said that they were looking into a way to get a document where structure is legit, but the content would be lorem ipsom

Laura Brady: Groundwood Books can donate chunks of picture books, graphic novels, MO books
… no DRM required

Will Awad: I have asked contact in Dubai to send sample of arabic book

Charles LaPierre: do we have any folks representing JP publishers on the business group who can submit samples of manga?

Wendy Reid: good idea. I’ll ask.

George Kerscher: you can mention that just portions of a book would be good enough

Wendy Reid: we probably just need like a sample that they would put up on a publisher website
… and i’ll also try to come up with some sort of quasi-legal text about how we will use
… so now that we have documentation, we just have to do the work of updating it
… i believe we’ve had people offer to work on best practices document. Ken and Rachel.

Ken Jones: i’ve just about finished the FXL reading order article
… wendyreid can you review for tone and language?
… basically built off of what I’ve already written. Using some images as well, assuming that is okay.

Wendy Reid: with a note we don’t have to be as strict with the tone

Rachel Osolen: i got some work done on accessible tables, and on accessibility metadata
… in the past we were talking about adding metadata for a11y, is that still something we’re thinking of?

Charles LaPierre: yes, in accessibilityFeatures. Probably not adding any hazards
… right now its just a wiki page based off of work Benetech did a while back

George Kerscher: accessModeSufficient = textual means all content can be consumed via text
… if content contains images with alt-text, does that pass?
… even if that book can’t be zoomed?

Charles LaPierre: from my understanding (based on Madeline Rothberg), as long as the alt-text has any text that is in the image, plus an extended desc of the image itself (if that is relevant), then accessMode sufficient = textual is okay
… for picture books, accessModeSufficient = visual is correct
… and then books that have a combination of text and images described by alt-text would be accessMode sufficient = visual, textual

Wendy Reid: so I’d assume that for many FXL it would be “visual, textual”

Ken Jones: so for “visual, textual”, does the entire content have to be consumable via each of them individually? or via the two modes in conjunction?

George Kerscher: i think the clearest way to do this is by having multiple accessMode sufficient statements

Wendy Reid: maybe we can run this section of our doc by madeline once we have it more complete
… but going back to Rachel_Osolen’s question, “yes” we can add metadata values
… if there is a compelling enough reason to do so

Rachel Osolen: so for now do I just say “we suggest that we expand the Schema”?

Wendy Reid: you can put that in an editorial note

George Kerscher: like gpellegrino said, there is AT that does magnification independent of content (e.g. make a picture biggest and pan around it)
… but normally when we think of reflow, we think of low-vision folks changing font size and making the visual adjustments of the text
… when we are making accessibility claims in our metadata, should we consider that zoomed panning acceptable?

Wendy Reid: i wonder if you should raise that with the epub a11y TF

George Kerscher: it’s a horrible reading experience
… and with the EUAA, that’s going to be a big big deal

Gregorio Pellegrino: it is the same user experience you have with PDFs, and the EUAA requires that the user can change font, size, and different design options to accommodate his disability

Ken Jones: I think that’s what accessModeSufficient = textual should mean
… i think we should just say that FXLs are normally visual documents, and only become textual if they can do all those things that gpellegrino said

Charles LaPierre: i wonder if we could have an FXL where the text was resizable

George Kerscher: there’s an accessibility feature about transformability?

Charles LaPierre: yes, so that’s the one that refers to resizing text, etc.

George Kerscher: so the combination of accessMode = textual + transformability, results in accessModeSufficient = textual?

Charles LaPierre: no, I don’t think transformability comes into accessMode sufficient

George Kerscher: but FXL is incompatible with transformability, yes?
… i think we can get reading order, and I think we can get AT to read the text, but unclear what to do about low-vision
… same thing with learning disabilities (need for personalized font, etc.)
… this is getting to the heart of what we are trying to figure out here
… (and the EUAA)

Rachel Osolen: this is why I think we should add more to the Schema for metadata
… to make it more clear what FXL can and can’t do

Ken Jones: what about metadata for the type of screen size that would be best for the content

Charles LaPierre: about transformability for dyslexia, there has been research that says that the reader just needs to get used to the font
… just being able to change the font, even if the size remains the same, would be helpful

George Kerscher: unless we can come up with a way to have a parallel HTML presentation I think we’re running up against a big problem here
… like with multiple renditions

Wendy Reid: if the HTML is good, the page itself is what you are referring to

Rachel Osolen: when I train people on how best to write code for accessibility, I stress proper HTML
… and the other thing I wanted to mention was proper use of colour

Brady Duga: Ken_Jones had an idea a while back about transforming FXL into reflow epub
… an algorithmic solution
… which would give us all the a11y benefits of reflow
… I think we eventually referred this idea to the CG
… if EUAA is on the horizon, then we have to start speccing that today
… but it seems like something like that would solve a lot of our problems

Ken Jones: i think we need to get away from the idea of changing the FXL, like changing the font, because i can’t see that ever becoming popular
… as soon as you go down that path you end up with badly designed FXLs that are probably still not very accessible

Wendy Reid: let me take this away, since it gets into questions of charter, etc.
… for those that have tasks about adding to/updating documentation, please share with the group via PR or email
… for the experiments, lay claim to the ones you are interested in and go wild
… ditto for the updates to the Daisy KB (and remember that Ken has already started that work with his outline)
… AOB?
… okay, thanks everyone, we’ll reconvene next week!