Screen Reader Friendly can be determined in two ways that should be described.

[GeorgeKerscher]

AccessModeSufficent with a value of textual is one way of determining that a screen reader will be able to access all the information. Graphics will need alt text and/or extended descriptions for textual to be applied.

Also, if the publication conforms to WCAG at the A or AA level, all graphical content must have a text equivalent.

Our guidelines should point out that the term screen reader friendly can use either or both of these metadata items.

[michael-n-cooper]

I removed the a11y-needs-resolution label as that triggers APA tracking of comments filed on W3C specifications. I think that is not the intent for issues in this repository.

[GeorgeKerscher]

We clearly address this in the techniques

[HadrienGardeur]

A number of metadata can be inferred from the fact that a publication conforms to WCAG A/AA.

We could either:

• mention conformance to WCAG as an alternate way of determining the presence of these features
• or address inferences from WCAG A/AA as its own section

Inferences are never without risk since a publisher/author could mistakenly indicate that a publication conforms to WCAG.

[rickj]

With the US legislation using WCAG 2.1 AA as the requirement, I would want more clarity in the techniques documents. The inferences @HadrienGardeur references are critical for those who are not immersed in the details to understand what is being claimed.

As an example: a strict reading of the variable setup for 'Supports Nonvisual Reading' would not direct an implementer to understand that a WCAG conformance claim would potentially affect this claim.

Perhaps we need a crosswalk of WCAG inferences, and when such a claim is made call those out in the respective areas? I'm not advocating for stating a factual claim, but something like "WCAG 2.1 AA Conformance has been claimed, which implies XXX. Users should verify this before creating any reliance on this implication"

[wareid]

Agreed on a crosswalk of WCAG inferences, but I do agree that if a publication is claiming WCAG 2.x AA conformance, screen reader friendly is a logical inference, especially considering how much of WCAG is devoted to ensuring good screen reader compatibility.

One helpful breakdown for publishers might be what in WCAG applies to book content. WCAG is huge and covers a lot of use cases, but many of them don't apply to ebooks, such as the authentication requirements.

We did a version of this kind of breakdown for the FXL accessibility doc, it wouldn't be too hard to extract and expand that content for reflow.

[clapierre]

One helpful breakdown for publishers might be what in WCAG applies to book content

DAISY's SMART tool which Benetech utilizes as part of our GCA certification program has already listed all relevant WCAG criteria that is relevant to EPUBs. You can also specify which version of WCAG 2.0, 2.1 or 2.2 you wish to use.

Matt did an amazing job setting this up for publishers to check their content manually and works in conjunction with ACE for any of the criteria which can be tested automatically.

[HadrienGardeur]

Based on data that I've presented last year in Frankfurt and Paris:

• most publications that claim WCAG compliance only do so in their accessibility summary and not using conformsTo
• publications that correctly claim to conform to WCAG usually include textual on its own in accessModeSufficient
• but most of these publications forget to also include displayTransformability in accessibilityFeature

Also worth pointing out that with over 2 million EPUB files available, I couldn't find a single publication that claimed conformance to level A, every single one of them targeted AA.

[clapierre]

Also worth pointing out that with over 2 million EPUB files available, I couldn't find a single publication that claimed conformance to level A, every single one of them targeted AA.

We have had a few GCA publishers publish a few books at the WCAG 2.0-A level because they failed to meet color contrast requirements due to some unique requirements enforced by the author which could not be adjusted and also mentioned why they are only Single A in the accessibilitySummary.

Others did the same because there were too many language shifts throughout the book and was too costly to tag each of them correctly.

[rickj]

We have 34,228 titles with an explicit WCAG level called out in the conformance metadata.

Rick Johnson | Co-Founder and Vice President of Solutions Engineering and Accessibility VitalSource Technologies, LLC get.vitalsource.com https://get.vitalsource.com/

From: Charles LaPierre @.> Date: Wednesday, September 11, 2024 at 10:50 AM To: w3c/publ-a11y @.> Cc: Johnson, Rick @.>, Comment @.> Subject: Re: [w3c/publ-a11y] Screen Reader Friendly can be determined in two ways that should be described. (Issue #123)

Also worth pointing out that with over 2 million EPUB files available, I couldn't find a single publication that claimed conformance to level A, every single one of them targeted AA.

We have had a few GCA publishers publish a few books at the WCAG 2.0-A level because they failed to meet color contrast requirements due to some unique requirements enforced by the author which could not be adjusted and also mentioned why they are only Single A in the accessibilitySummary.

Others did the same because there were too many language shifts throughout the book and was too costly to tag each of them correctly.

— Reply to this email directly, view it on GitHubhttps://github.com/w3c/publ-a11y/issues/123#issuecomment-2343899907, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAJQELSFAH7WBTWEOIUIT5TZWBKDLAVCNFSM6AAAAABNXJASKSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNBTHA4TSOJQG4. You are receiving this because you commented.Message ID: @.***>

[avneeshsingh]

The topic was discussed in Sept 12, 2024 meeting:

AvneeshSingh: shall we infer things from a conformance?

gautierchomel: for me the rules for inference are another thing from the guide … mainly because this inference can be done on publishers' side, but not by distributor … we should not recommend inference if they're not managed by publishers

AvneeshSingh: there's one inference for figuring out accessibility metadata from analysis of the content, it is a different inference when it is from conformance metadata.

George: with visual adjustment, audiobook will fail, same for PDFs probably.

mgarrish: accessibility features are intended to describe more precisely than WCAG levels. I'm not sure how much we can really infer from WCAG.

Simon_M: I don't fully trust content creators claiming WCAG compliance. Certain types of media are inexactly mapped.

Simon_M: E.g. A publisher could think that the existence of Liquid Mode in Adobe Acrobat makes their content reflowable

George: +1

Madeleine: the risk has been exposed. Drawing inferences may disappoint users.

CharlesL: agree, we've seen misuse of compliance at Benetech certification. Without certification from a tiers, we cannot really trust a conformity claim. There is a need to check inside the book to establish features.

AvneeshSingh: I think we agree that we should not go this way. Will update the issue.