Industry experts from scholarly and educational publishing provide insight and advice regarding the requirements of the European Accessibility Act (EAA) and ADA Title II. They cover the measures publishers are taking to update both their frontlist workflows and their backlist content to make their books sufficiently accessible to comply with those requirements, including strategies for prioritization, dealing with image descriptions, and providing proper metadata.

Our experts discuss their accessibility efforts, both internally and with suppliers, to sufficiently conform to those two key regulations. The panelists discuss common misconceptions; the roles of authors, editors, designers, and suppliers; and practical measures that can be taken in the short term while more thoroughgoing improvements to their workflows can be implemented in the longer term.

Presenters:

• Bill Kasdorf – Principal at Kasdorf & Associates, LLC
• Stacy Scott – Accessibility Lead at Taylor and Francis
• Rachel Comerford – Senior Director of Accessibility at Macmillan Learning
• Joshua Routh – Director of Hosting Solutions at HighWire Press

Q&A

Q1. My main question is regarding what we need to do for ALT text and accessible PDFs for the journal titles and preprint servers.  I understand that EAA mainly focuses on paid products.  (Ignore if this will be covered in the presentations.)

Answer:

Bill: You’re right that the EAA focuses on ecommerce, but if the journals or their articles (including preprints) are subscription products then my advice is that they apply. But note that in any case they will be an issue for ADA Title II, which applies to web content and web apps. If your journal titles and preprints are accessed in public libraries, schools, or universities, you’ll be under pressure to make them accessible.

Q2. Do these accessibility changes required to be change to historical content?

Answer:

Josh: Ultimately, yes, but the EAA timeline for the backlist is more generous than for new content. For front list it is June this year. For backlist is is 2030.

Bill: Most accessibility experts I consult (and including me) advise clients not to count on that five-year extension for backlist. We need to see how that is interpreted in the final implementations by the 27 member states. Virtually all of my clients are working on getting backlist accessible (at least making progress on it, usually based on prioritization). And it applies to ADA Title II as well, which is April 2026.

Q3. Related 2 part question, for HighWire (or others!): Is it possible (and straightforward) to get some kind of usage information on accessible content, i.e. which articles, books, etc. or how much content overall is accessed using assistive technologies? 2nd part – if not, is it possible to know which assistive technologies/tools are being used to access content and ideally how often? Thank you!

Answer:

Josh: Hi. To the best of my knowledge, there is no technical way of knowing what proportion of your content is being accessed using accessibility software such as JAWS, as they do not have example a unique user agent.

Rachel: Agreed. Additionally, this information is difficult to gather with privacy protections. There is a regular WebAim survey about screen reader usage popularity that will give you an idea of this assistive technology specifically, but it’s worth noting that screen readers are not the only assistive technology. https://webaim.org/projects/screenreadersurvey10/

Other assistive technology includes simply using a keyboard to navigate… It’s fairly safe to assume that the majority of your readers are using one of those (but not all!)

Q4. What is ARIA?

Answer:

Rachel: AIRA, or WAI-ARIA is the Web Accessibility Initiative – Accessible Rich Internet Applications. Its a spec that the W3C published (world wide web consortium) and helps web pages communicate with assistive technology about how to navigate the page when standard HTML isn’t enough.

Q5. Are authors trained to supply the best “extended descriptions?”
What’s the best practice for scholarly journals to supply extended descriptions?

Answer:

Bill: This is very publisher-specific. Most publishers provide some basic guidelines to authors, knowing that an editor will refine the author’s image descriptions. But some publishers provide quite extensive guidelines, for use not just by authors but their own editorial staff and freelancers. One of my clients has done a really fantastic job of that, with excellent instructions, guidelines, and examples.

Q6. Given that the EAA guidelines apply to PDFs as well as webpages, why are software businesses like Adobe not making these changes to their programs at the source and issuing a blanket software update, rather than thousands of publishers having to liaise with third parties to make their PDFs compliant?

Answer:

Bill: Actually, Adobe has made great progress recently (especially thanks to consulting with experts like Gregorio Pellegrino of Italy’s LIA) on building functionality into the latest versions of InDesign to enable the generation of accessible EPUBs. It even provides drop-downs to facilitate proper metadata–the required properties and the controlled vocabularies associated with them. So for publications created with InDesign, it may soon be easier to make an accessible EPUB than an accessible PDF. PDFs can’t be as accessible as EPUBs, as several of us stressed in the webinar. But Acrobat Pro has a very easy-to-use feature for testing accessibility, including showing what image descriptions are present for every image in the file, with an interface that makes it easy to apply or fix image descriptions. Checking reading order is easy too (although fixing it isn’t).

Q7. Is HighWire going to verify if the publisher’s sites are conforming to EAA, ADA, and WCAG requirements?

Answer:

Josh: Hi KJ. The HighWire platform conforms to EAA and ADA requirements by virtue of being WCAG 2.1 AA compliant. We can share a VPAT with you showing the tests we’ve done to confirm that, which we continue to update regularly.

Saying that, although we can verify that the interface is compliant, we also need the publisher’s input for the content itself. For example, we do accept images without alt text if that is what the publisher shares with us. We do not as standard review every page of every site to check all alt text has been provided. We’re happy to work with you on an approach for that though.

Q8. I understand the focus of accessibility is XHTML and ePub. PDF will never be as accessible. However, when will Adobe support short and extended descriptions? We flow alt text and long descriptions from XML to PDF, and that is problematic when Adobe supports only one image description field.

Answer:

Bill: I can’t speculate about Adobe’s plans. At this point, it’s best to just do what would be in an extended description if one is needed. If all that’s needed is a short plain text description as would be in alt text, then that’s all you need to put in that field. Note that Adobe doesn’t use the term “alt text.” That’s appropriate because technically that’s an HTML term. A good practice is to put yourself in the position of the print-disabled person: what information do they need about that image? That’s what you should provide.

Q9. We were told that as long as our html is compliant, our PDFs do not have to. Is this true?

Answer:

Bill: I’m not a lawyer, but that’s how I read the Directive. A digital publication has to be available in an accessible format. That doesn’t mean every digital format has to be accessible. But again IANAL, we’ll have to see what the 27 Member States implement.

Q10. Is frontlist and backlist the same as frontend and backend?

Answer:

Bill: I’m not sure what you mean by this. Frontlist is new books or newly published articles; backlist is previously published books or articles that are still available.

Q11. What is Ace and Thorium?

Answer:

Rachel: ACE is an epub accessibility checker hosted by the DAISY Consortium. Thorium is an open source e-reader that you can use to open epubs.

Q12. Bill, you mentioned that the UK has opted to adopt the EAA alongside the 27 Member States. Has there been an announcement about this and is there a link that can be shared to it?

Answer:

Bill: While that is my understanding, I will defer to Stacy on this, since she’s in the UK.

Q13. Are all authors being sufficiently advised to provide descriptions during submission? Or is the focus more on publishers and vendors creating descriptions for frontlists? The challenge with the latter is that there may be an increase in time to publish, given the author review cycle of accessible description.

Answer:

Bill: This varies widely by publisher, and by the type of content. In some sectors, images are commissioned from illustrators and the authors don’t originate them.

Q14. I work for a very small society publisher whose audience consists exclusively of surgeons – a group who, by definition of their occupation, cannot be visually impaired to the point that they need assistive technology to read our webpages and PDFs. Yet our small team is required to invest time, money and effort into making our web content EAA-compliant. My question, for our company and others that cater to similarly niche audiences, is why can’t we just continue with our current policy of providing accessible versions of our content if/when an individual requests them?

Answer:

Rachel: I would emphasize that while people with visual impairments benefit from accessibility work, they are not the sole audience. Surgeons can have learning disabilities like dyslexia or be neurodivergent. They can have hearing impairments. They can have physical disabilities. The solutions we are talking about benefit all of those audiences as well as the temporarily able-bodied audience… for example those who want to listen to an article while they commute. The goal is for all users to get access at once rather than telling some audiences they will need to wait.

Q15. With “Meaningful” being somewhat subjective, what are more objective ways to know whether an image description is meaningful?

Answer:

Bill: The guidance is that the description should convey to a print-disabled user what the image conveys to a sighted user; but it should not repeat what is already in the text or the caption.

Rachel: One way that you can experiment with quality is to grab a sample of the alt text that was authored, without the image, and ask a seeing colleague what the image looks like based on the description. Can they tell/show you what was being described? Was there too much info and it got confusing? Too little and there were gaps? Irrelevant content given the context the image was used in?

Q16. Is installing UserWay on our journal pages a partial solution? A non-solution?

Answer:

Rachel: I recommend doing thorough research on any overlay solution before implementing it on your platform.  Without commenting on any particular service, and there are many, there is a lot of commentary in the accessibility community about overlay use that is worth reading to better understand how they do and do not work.

Q17. How far does backlist go? What is we have content from the 1800s?

Answer:

Bill: There is a “disproportionate burden” exception in the EAA for cases where the effort and cost of making something accessible can’t be justified by the expected revenue to be obtained from it. (It’s more complicated than that; the Directive spells out how that’s determined, and it has to be revisited every three years.) So if your oldest content brings in virtually no income, that exception might apply. (Note, though, that the ADA Title II might still apply.)

Q18. T&F epub query – does an EPUB file cover each issue or is it done article by article?

Answer:

Stacy: This very much depends on your set up and workflow. For T&F there’s one article for each EPUB. We create EPUBs based on each article’s XML. Any content issues identified would affect that one article only. Things like making headers or links easier to read is styling and Atypon (T&F’s platform host for journals) would fix their stylesheet and then reprocess all EPUBs to make those changes live.

Q19. Can the speakers share a bit more about accessibility metadata? Does it need to be third party certified? On arXiv we have 1000+ submissions each day and we are aiming for ‘better’ rather than ‘perfect’. Each submission will not always pass all criteria.

Answer:

Bill: Hi, Shamsi! I’d suggest you consult the EPUB Accessibility 1.1 Recommendation from the W3C. It explicitly spells out which properties are required and which are optional.

Rachel: Third party certification is not a requirement. I would also suggest looking into quick wins – if you have a consistent workflow for producing content then some elements of the accessibility metadata will never, or will rarely, change.

Q20. Bill, Benetech has recommended two practices that the old DAISY Ace disapproved. One was the use of dc:pagesource, which I think you spoke to. The other is the encoding of pagebreaks. Benetech (in Technical Bulletin 018 of March 2025) recommended the following paradigm in ePub3:

<span epub:type=”pagebreak” id=”page5″ role=”doc-pagebreak”
aria-labelledby=”pg5″></span>
<span aria-hidden=”true” id=”pg5″>5</span>

Will the new DAISY Ace forgive both of those Benetech recommendations?

Answer:

Bill: I would consider this ideal. The reason is that reading systems may still be looking for the epub:type attribute. You’re doing both, which is a good belt-and-suspenders strategy. I would not cite these as wrong to a client.

Q21. What is the point of publishing journal articles as e-pubs vs. accessible pdfs?

Answer:

Stacy: EPUB is the clear choice for accessibility, meeting EAA standards by allowing users to adjust fonts, spacing, and colours—something PDFs can’t fully support. Built on the Open Web Platform (HTML, CSS, XML), EPUB integrates the accessibility features of the DAISY format and is the global standard for e-books, accepted by all major distributors.

Q22. Is there advice on how to tackle this if your workforce is less than 1 FTE?

Answer:

Josh: We get this question a lot – you’re not alone! In the final section of the webinar I listed 10 points which could be used by organisations with a very small team. Almost all of them are variations on activities done by larger publishers at scale, but with a focus on pragmatism. None of the tasks should take more than a few hours, and some take only a few minutes.

Stacy: Not to be considered legal advice, but my strong understanding is that Microenterprises which employ fewer than 10 people or have an annual turnover of less than 2 million euros are exempt from the EAA.

Q23. Is HighWire offering audits on our content?

Answer:

Josh: At present we’re not offering a service to check accessibility on non-HighWire hosted content. We’re certainly happy to have that conversation on content hosted with us.

Q24. Can anyone comment on the EAA+ADA regulations for video content? What are TF doing to support this for journal content and what deadline are you targeting?

Answer:

Stacy: T&F are looking into what is needed to make our videos fully inclusive. This includes, closed captioning, transcripts and audio description. We are in the process of discovery, ascertaining how many videos we have in need of further enhancements to make them fully accessible. We do not yet have a definitive timeline for this.

Q25. My answer to the “if it is rarely used, is it worth it?” question, is if it is updated to be as accessible as possible, maybe it will be used.

Answer:

Rachel: Thanks! That’s a great point. I was thinking more about outdated content that contributes to volume but not quality of content. Sometimes we keep this around ‘just in case’ but I think updating content to be accessible is also a great opportunity to review content quality in general and decide if it’s worth keeping.

Q26. Regarding alt text: Do we need to define acronyms in an alt text again if they are already defined in the text or figure legend?

Answer:

Bill: No. A cardinal rule of image descriptions is that they should NOT repeat what is in the text or legend or caption. Plus, aren’t they text, rather than images, anyway?

Q27. In the case of e-commerce invoices in PDF format, what can you suggest as an alternative to provide that meets EAA requirements? And how to champion another format when people are concerned about breaking away from what’s considered the expected format for an invoice.

Answer:

Stacy:  Not to be considered legal advice, but my strong understanding is you can still make a PDF accessible and as long as your invoices are made accessible, I don’t think this should cause a problem. The key focus of the e-commerce aspect to the EAA must include all digital interaction, including card payments, billing etc.

Q28. Do we have potentially greater liability in Europe because we have pay-per-view articles available?

Answer:

Stacy: Not to be considered legal advice, but my strong understanding is that if these are journal articles, then the consensus is that journals are not covered under the EAA.

Q29. Bill, I’d agree with you that Benetech’s recommended page-break mark-up is optimal. But DAISY Ace reports it as DAISY Ace reports that as an error – “pagebreak-label”. Does the new DAISY version account for the new Benetech recommendation?

Answer:

Bill: I can’t speak for DAISY but if you raise this issue with them they do pay attention. It appears that their software is recognizing the epub:type and triggering an error even though you also have the recommended doc-pagebreak.

Q30. On PDF accessibility, we have PDF only workflows…. can we still be compliant to the EAA with only PDF? Or will we be liable for those?

Answer:

Bill: I am not a lawyer, and can only speculate, but if your PDFs conform to PDF/UA I can’t imagine that wouldn’t be sufficient; that’s the current standard for an accessible PDF.

Q31. We have legacy articles form 1917 with the abstract in html and the rest is a scanned in PDF–do we have to ultimately make those accessible?

Answer:

Bill: See my answer to a similar question above.

Q32. Should we switch from PDF to ePubs going forward for our journal publishing?

Answer:

Stacy: EPUB is the clear choice for accessibility, meeting EAA standards by allowing users to adjust fonts, spacing, and colours—something PDFs can’t fully support. Built on the Open Web Platform (HTML, CSS, XML), EPUB integrates the accessibility features of the DAISY format and is the global standard for e-books, accepted by all major distributors.

EPUB’s flexibility benefits both people with disabilities and the general public. It works across mainstream and assistive reading tools, supports automatic text-to-speech, electronic braille, and customizable layouts, and includes semantic tagging for blind readers.

In contrast, PDFs are tied to print conventions and lack essential accessibility features like reflowability and font customization, making them a poor fit for EAA compliance, visually impaired readers, and small screens. They’re also widely rejected by e-book retailers—publishers should prioritize EPUB instead.

There are many resources to read why EPUB is superior to PDF from an accessibility perspective, please visit daisy.org to learn more, and read T&F’s view here:

https://insights.taylorandfrancis.com/social-justice/epub

Q33. How will these standards be enforced? Will we be penalized as journal publishers?

Answer:

It depends on your location – in the EU, each member state has its own enforcement plan which can include fines or in extreme cases, prison sentences. This article from Level Access summarizes a handful of them: https://www.levelaccess.com/blog/penalties-for-eaa-non-compliance/ but there are many more.