Coloured pattern Coloured pattern
Read all the news

DPUB ARIA 1.1 released. How to improve EPUB semantics

Research and development

The World Wide Web Consortium (W3C), the international community that develops, maintains and manages the specifications of web-related standards, has released the First Public Working Draft of the Digital Publishing WAI-ARIA Module 1.1.

The DPUB ARIA specification can be defined as a semantic dictionary that includes a taxonomy of attributes, called “roles”, that can be added to HTML tags to enhance their semantics. The roles, states, properties and values of the module are specific to the digital publishing industry and allow publishers and creators of digital content to convey information about the structural semantic of a publication, for example an ebook in the EPUB format and long-form web documents, to users, including assistive technology users. The taxonomy includes roles such as “doc-abstract”, “doc-appendix”, “doc-chapter”, “doc-prologue” (among others): applied to HTML elements, they identify the function performed by the textual components of the publication, allowing a section to be marked as an abstract, or as a chapter, or as a prologue, thus solving a problem of HTML not having adequate native tags for these elements.

Although the latest version of HTML, HTML5, contains some semantic tags, i.e. tags that convey information about their content and its function within the page (like <header>, <footer>, <section>…), the web and its languages are not designed for the publication of long-form, structured documents, and for now there are no specific tags to indicate that a paragraph of text is, let’s say, an abstract, a colophon or a footnote.

The DPUB ARIA specification aims to fill this gap. Unlike common web pages or HTML documents, publications are characterised by a semantic structure that has been defined over time through evolutions and modifications and reproducing this peculiar semantic structure in a digital publication enables a semantic navigation of the document and improves the reading experience of assistive technology users at the same time.

Indeed, embedding DPUB ARIA roles (i.e. semantic metadata) enhance the accessibility of digital publications for users with disabilities, since assistive technologies can inform the users about the content structural division and the behaviour of web-application widgets.

Major changes

The public draft defines which roles have been added to the WAI-ARIA role taxonomy and describes the characteristics and properties of all roles.

The primary aim of the revision was to address some bugs discovered since the release of the original module. Among fixes mostly minor in nature, a couple of significant changes are proposed in this revision:

  • The doc-endnote and doc-bilbioentry roles are being deprecated due to their incompatibility with the core ARIA inheritance model for roles
  • The introduction of doc-pageheader and doc-pagefooter for identifying running headers and footers, so readers can choose whether to suppress the repetitive content they contain.

What is the difference between DPUB ARIA and epub:type?

An attempt to introduce semantics in EPUB files was already made by the IDPF, the International Digital Publishing Forum, with the implementation of the EPUB Structural Semantic Vocabulary and the epub:type attribute. This vocabulary defines a set of properties relating to the description of structural semantics of written works and was specifically developed for EPUB documents.

Although the epub:type attribute and the DPUB ARIA roles may seem very similar, their purposes and applications are quite different.

The epub:type attribute is used to express domain-specific semantics in EPUB Content Documents and Media Overlay Documents, using the structural information it carries complementing the underlying vocabulary. Currently the default vocabulary for the epub:type attribute is the EPUB 3 Structural Semantics Vocabulary, which is part of the EPUB3 standard.

Applying the epub:type attribute to an element does not change its nature for assistive technology, nor does it enhance the accessibility of the content. Instead, it just provides hints about the purpose of an element and enriches content for use in publishing workflows and for author-defined purposes. In brief, the epub:type attribute is only intended for publishing semantics and Reading System enhancement, it does not impact accessibility nor the behavior of assistive technology.

The DPUB ARIA, on the contrary, is an extension of the WAI-ARIA, a technical specification aiming at improving web accessibility and interoperability with assistive technologies.

Currently, the two technologies coexist.

How to include DPUB ARIA roles in EPUB publications?

Unfortunately, the support of DPUB ARIA roles by authoring tools is still very limited. InDesign, for example, allows to assign a semantic to some parts of the publication, which is translated into the epub:type and its reference value in the semantic dictionary, but does not support DPUB ARIA roles.

It is to be expected that over time authoring tools will increasingly support these roles, enabling the creation of more semantically structured digital publications that are more accessible to users of assistive technologies.

Is DPUB ARIA supported by assistive technologies?

This question is asked by many content creators, but even within the accessible digital publishing community it is not clear what support assistive technologies offer to metadata expressed with DPUB ARIA. Starting from this cue, a mapping work has begun within the working groups of the DAISY Consortium, with the support of Fondazione LIA.

This is a work in progress: we started by mapping the support of DPUB ARIA within browsers (since browser engines are the basis of many reading solutions) on different platforms and then checking the interaction between browsers and assistive technologies (screen readers).

Initial results show that there is still work to be done, but one combination amazed us: using Chrome on Android, with TalkBack assistive technology most of the DPUB ARIA roles are recognized and communicated to the user.

Now the work is to raise awareness among the developers of the different platforms so that they increase the implementation both on the browser side and on the assistive technology side.


The document was published by the Accessible Rich Internet Applications Working Group and, although is intended to become a W3C Recommendation, is still a work in progress, meaning it could be updated, replaced or obsoleted by other documents at any time.

You can see the progress of the work on the public repository: https://github.com/daisy/transitiontoepub