The Accessible EPUB from InDesign Expert Group has been engaging with the InDesign engineers since February 2022 on a series of issues relating to the EPUB export from layout files in InDesign. Richard Orme (DAISY) facilitates the group, Gregorio Pellegrino (Fondazione LIA) is our technical lead, with contributions from Jonas Lillqvist (CELIA) and Laura Brady (eBound/Accessible Books Consortium).

The Expert Group collaborates with the Adobe team, providing them with detailed information to improve the export of accessible EPUBs from InDesign. The goal is to reduce the manual and time consuming efforts that are currently required by publishers to ensure content is accessible at this stage of their workflow.
Prodigious and relevant progress has been made, the most notable of which is the inclusion of page navigation and a full suite of accessibility metadata options as of February 2024’s 19.4 release.

In mid-July, a further step was achieved with the version 19.5 release. A few bugs were fixed to improve the way that the pagelist exports. The <div> and <span> tags marking page breaks are now void elements that are self-closing.

This markup:

<div id="page139" role="doc-pagebreak" aria-label="139" epub:type="pagebreak" ></div>

Now looks like this:

<div id="page139" role="doc-pagebreak" aria-label="139" epub:type="pagebreak" />

For content creators, the most significant development in 19.5 is the more robust population of the <title> field in the HTML files. That element was previously filled in with the iterated name of the InDesign file — so, InDesignFile-1, InDesignFile-2, etc. This was a misuse of that field which no doubt created considerable confusion for print-disabled readers when read by assistive technologies.

The InDesign export now looks for the text from the top-level header in that HTML file (tags h1, h2, h3, h4, h5, h6), duplicating it into that element. This is a significant improvement, saving many ebook developers a step in their remediation of their EPUB export. Where a top-level header exists, this new feature works quite nicely. There are a few tiny kinks to work out for HTML files without a header or that contain only an image. As with much of this work, it is a process.

There are many other nice fixes on the horizon. Watch this space for news!