W3C

– DRAFT –
FXL Accessibility Task Force Meeting 30 May 2023

30 May 2023

Attendees

Present
duga, gautier, gpellegrino, jeffrey_griggs, JonasLillqvist, Miia, Naomi, shadi, wendyreid
Regrets
-
Chair
wendyreid
Scribe
gpellegrino, wendyreid

Meeting minutes

<wendyreid> date: 2023-05-30

wendyreid: this is our first meeting in the Publishing CG
… we moved here from the WG, since this work is more on incubation
… there'll be a new WG for maintenance of EPUB3

duga: I'm Brady I work in Google, I'm in EPUB space since many years

gautier: I'm Gautier I work for EDRLab

JonasLillqvist: I'm Jonas, I work at CELIA, the DAISY library in Finland

shadi: I'm Shadi from Amazon, I was a W3C staff

jeffrey_griggs: I'm Jeffrey from Penguin Random House

Miia: I'm Miia from Celia

wendyreid: I'm Wendy from Rakuten Kobo, I've been the chair of the EPUB 3 WG, now I chair this group

wendyreid: First question: what do we want to focus on?
… we started with guidelines and ideas for making fixed layout more accessible
… one is the idea to being able to switch from FXL content to the same content in reflowable

<wendyreid> https://w3c.github.io/publishingcg/TaskForces/fxl-a11y/index.html

<wendyreid> https://w3c.github.io/epub-specs/epub33/fxl-a11y/

gpellegrino: From experience as LIA, in the last month we assessed an accessible FXL book, we sent to some blind testers, and they asked why?
… experience for them was no different than any other book
… maybe we're not so far away
… we do need to explore some ways for low vision users to get the content
… fxl content for low-vision users is tricky
… maybe that work is not on the side of content creators, but on the side of reading solutions

gpellegrino: Another related thing, in WCAG as far as I know, it is required to be able to zoom fxl content by 200% or 400% which is the way PDFs can be compliant
… since we know PDF reflow is limited
… we know most reading solutions limit zoom for FXL, especially on desktop
… maybe having some guidelines or suggestions for RS developers may be useful

gautier: from the RS development side, we've seen different projects in the past months
… and been asked different implementations
… we can't develop an implementation for every proof of concept
… I think that before developing, we need to get a common way to resolve the issue
… so to be sure that our effort it useful
… for sure having the zoom is fine, but it is the minim level
… the problems we see in the EPUBs is that they're mainly created by InDesign

wendyreid: for sure there is room for guidelines for reading systems
… zoom is important, but is not sufficient
… think about mobile readers
… but also think about FXLs made only of images
… there we cannot extract the text

gpellegrino: My first idea was something like the guide from the PDF association for getting HTML out of PDFs, a standardized approach
… that enables every content creator to create a PDF that can be made reflowable
… I think this might be one approach
… accessibility means quality
… a subset of building FXL content, don't focus necessarily on tools
… what do we need from HTML to have better reading experiences
… I know many FXL titles are image-only
… those FXL aren't accessible at all
… it can be done, but in this TF we can try to raise the bar
… if you want to make a FXL out of images, fine, but if you want it to be accessible, do this

https://pdfa.org/resource/deriving-html-from-pdf/

wendyreid: we can give to the users all the information, but then is on their side to decide to do something or not
… like making a content accessible or not

wendyreid: so for the moment we have: guidelines accessible FXL reading systems, guidelines for content creators to enable them to make great FXL books that can be manipulated by reading solutions (making them reflowable)
… like PDF association

Naomi: something very useful may be a mapping between a11y metadata and code examples
… so we can start from something

gpellegrino: I think most of us here knows Lars from Colibrio, Lars thinks that SVG can be a game changer for making accessible FXL with complex layouts, using ARIA and SVG
… I haven't tested it yet
… not sure if it's supported by assistive technologies, but a better way to export a book from authoring tools
… pixel perfect without losing semantivs
… maybe a research project to check that
… we focused mainly on HTML/image formats
… but SVG is also supported by the standards

Naomi: we've never done SVG in spine
… but we did embed SVGs in HTML
… but not all reading systems support it
… so this may block content creators
… I think it'll need more research

duga: we implemented this at the begin of the pandemic
… we used SVG and used ARIA plus Media Overlays
… it was pixel perfect
… even if text is a drawing, you can group it and make it acts like text
… the Media Overlay works perfectly
… we didn't continue with this project
… we has issue in grouping words and lines

wendyreid: one question is: what is the time on production side?

Naomi: I have code that converts text on a path in InDesign in SVG
… it was not simple, but can be done
… I think InDesign is not that interested in exporting SVGs

duga: we used Google slides as sources, where you can export SVGs

gpellegrino: I think you can get quality SVG out of PDFs
… it's a strange workflow, but if you get a good PDF, you can then convert to SVG, then add ARIA

wendyreid: it worths doing research on this
… from EPUB tests we know that SVG is well supported by different reading systems

Naomi: we had problems with text on a path
… SVG in general may work, but specific features may not

wendyreid: region base navigation is a thing
… it's not well supported right now
… I think it is important to investigate the needs of users
… link cognitive disabilities users, they may benefit for narrated books

gpellegrino: Two things, on the issue side, I could open an issue for SVG investigation just ot have something to follow.
… there is a lot of hype right now around accessible FXL, at the same time there is some misinformation out there
… if accessible FXL is a new specification
… in my opinion, accessible FXL is just a FXL EPUB3 following WCAG
… there are different interpretations
… media overlays but without semantics

gautier: I think the document linked by Wendy is a reference point as state of the art
… maybe from there we can link to different solutions from there
… so that then we can have a short list of the proposals
… because it is difficult for reading systems developers to implement them all

wendyreid: Ok, we can open different issues
… and then I can add a paragraph on the note, telling what is an accessible FXL
… I'll make a PR on that
… if you have new ideas on this, please open an issue in the CG repository
… we have different task forces in the CG

https://github.com/w3c/publishingcg/issues

wendyreid: se you all in 4 weeks

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

Diagnostics

Succeeded: s/SCG/SVG/

No scribenick or scribe found. Guessed: gpellegrino

All speakers: duga, gautier, gpellegrino, jeffrey_griggs, JonasLillqvist, Miia, Naomi, shadi, wendyreid

Active on IRC: duga, gautier, gpellegrino, jeffrey_griggs, Naomi, shadi, wendyreid