Overview

Reading a digital publication is a very personal experience. For most people this is routine, and little consideration is given to how the title was obtained before it is read. Users may go to a bookstore or library, search for the title to purchase online, or have the title selected for them by an instructor for a class.

Now consider that the person is blind and relies on assistive technology. The user needs that technology to assist them in the purchase process as well as to read the e-book. The person may wonder: will the screen reader work with this title; are there image descriptions that will be spoken to describe these images; are there page numbers which are accessible; is the reading order correct so a caution will be announced before reading the paragraph which could be dangerous? All of these accessibility concerns are potential issues consumers have when trying to purchase and ultimately read a digital publication in any format.

The good news is more and more publishers are creating digital publications that are Born Accessible (i.e., accessible from the outset, not fixed later) and getting the accessibility validation or audit done by independent organizations.

These guidelines help those who wish to render accessibility metadata directly to users understand how to represent the accessibility claims inherent in machine-readable accessibility metadata in a user-friendly User Interface / User Experience (UI/UX). This document targets implementers such as bookstores, libraries, retailers, distributors etc. Content creators will benefit from reading these guidelines and are encouraged to follow EPUB Accessibility 1.1 Conformance and Discoverability Requirements section and its techniques.

Metadata found either inside a digital publication or in a corresponding external record may have important accessibility claims that help end users find and determine if the publication can meet their specific accessibility needs.

This accessibility metadata uses controlled vocabularies to allow it to be extracted and displayed uniformly across different publications and localized to different user interface languages. The one exception is the accessibility summary, which allows accessibility statements that are unique to a publication and that adds information not covered by other metadata entries.

One important aspect is that the role of the Accessibility Summary metadata has changed in the latest version of the EPUB Accessibility specification, so a more in-depth analysis in the Accessibility summary section is recommended.

This document offers guidance on how to aggregate and display claims inherent in metadata to end users; these are not strict guidelines, but suggestions for providing a consistent experience for users through different portals. Different implementers may choose to implement these guidelines in a slightly different way. Some examples can be seen in the Implementations section of the document.

Metadata processing

The following diagram depicts how these guidelines relate to the format-specific techniques in the process of receiving and displaying accessibility metadata.

The diagram categorizes two ways that metadata accompanies a publication. In the first are digital publication formats that directly embed accessibility metadata (EPUB and PDF). In the second are external metadata record formats (ONIX and MARC) that accompany a digital publication as it moves through the supply chain.

In some cases, a digital publication may include both internal and external metadata (e.g., an EPUB could have accessibility metadata in it package document and also be provided to a vendor with an ONIX record). In these cases, vendors and reading system developers determine for themselves which set of metadata they will use to display to users.

The next level of the diagram depicts the encoding standards expected for each format. This guide assumes that the metadata conforms to a recognized standard, otherwise it would be difficult to anticipate and process the incoming information for users:

• For EPUB, the EPUB Accessibility standard [[epub-a11y-11]] defines required and recommended metadata for accessible publications. Guidance on applying this metadata is found in the EPUB Accessibility techniques [[epub-a11y-tech-11]].
• For ONIX, the accessibility characteristics of the corresponding digital publication are expressed using code lists [[onix]], and guidance on applying these values is defined in the application note Providing accessibility metadata in ONIX (advanced).
• For PDF, the PDF/UA standard [[iso14289-1]][[iso14289-2]] and Well-Tagged PDF guide define how to describe accessibility metadata.

Knowing how the metadata is encoded and how it is expressed, the next stage of processing is to use the algorithms defined in the respective techniques documents to discover and translate the information into human-readable statements. These documents are intended primarily for developers, to help them with the specific processing of the metadata markup grammars.

Finally, this document is shown at the last stage of processing, as it defines the purpose of each piece of metadata in more detail and prioritizes the display for readers.

Display techniques

To assist developers in implementing these guidelines, in-depth notes are available to explain how to extract information from publishing industry metadata standards.

At the time of publishing this document the available techniques for metadata standards are:

• Display Techniques for EPUB Accessibility Metadata 2.0

• Display Techniques for ONIX Accessibility Metadata 2.0

Terminology

There are several terms used in these guidelines that should be defined for clarity:

digital publication

The term digital publication is used in this document to refer to publications produced in any number of digital formats. Digital publications are not limited to books, but encompass any written, visual, or audio work distributed and read in digital form.

Some examples of digital publications include ebooks, audiobooks, manga, comic books, journals, digital textbooks, picture books, and children's picture books with accompanying audio. The formats they come in include EPUB, PDF, and Digital Talking Books (DTB).

dynamic braille

The term dynamic braille is used to denote content that is generated as braille on the fly, as opposed to preformatted digital braille formats. This dynamic rendering of content is sometimes referred to as electronic braille or refreshable braille.

Dynamic braille is typically rendered on a separate device from the reading system, one with pop-up pins to present the braille on a tactile screen. These devices, commonly referred to as refreshable braille displays, can be attached to a personal computer, or they may be a self-contained multipurpose note taker with a refreshable braille display.

read aloud speech

The term read aloud speech is used to denote content that is generated into synthetic speech on the fly, as opposed to prerecorded narration. Read aloud functionality is often a feature of reading systems, but can be provided by a separate assistive technology.

reading system

All digital publications require a reading system to present the publication to the end user. Reading Systems may be Apps that run on a smart phone or tablet. There are Reading Systems which are applications that run on personal computers. There are also Reading Systems that are integrated in to dedicated devices devoted to a single purpose, for presenting a publication. There are even skills that run on smart speakers that can be considered Reading Systems.

Accessibility section

Alternative statements

Although the display statements listed in this document and in the techniques documents are recommended for use, and have been translated to support localization, implementors may prefer to use different wording in some situations.

This document does not restrict implementers from using alternative phrasing, however any deviations should be done with caution. The phrases defined in this document resulted from much discussion among accessibility professionals as well as from user feedback.

If the cause of alternative wording is due to translation issues, implementers are encouraged to techniques documents for correction.

Non-accessibility metadata

Not all metadata used in determining the accessibility of a publication is strictly categorized as accessibility metadata. The product details, for example, provide important information about the usability of an ebook in relation to specific user needs.

In particular, the following information should always be displayed:

• File format (EPUB 2 or 3, PDF, MP3, Audiobook, etc.) — The file format gives a strong indication of accessibility: a PDF does not allow for typography modification, EPUB 2 is deprecated, an EPUB 3 support page navigation and better structural semantics; an MP3 format audiobook will be less structured than an Audiobook, etc.
• Protection measure or lack thereof — The protection measure may block assistive technologies such as screen readers. In addition, many specific eReading devices such as DAISY readers or Braille note takers are not equipped to read encrypted files.
• Name of the publishing house — The name of the publishing house can highlight the efforts it has made in terms of accessibility.
• Main language of the content — The main language of the content enables readers to be confident that they will be able to read with their Assistive Technology that uses synthesized voice or the corresponding braille translation table for the language.

This metadata does not have to be included as part of the accessibility section; it only needs to be made available to users in the metadata display for a publication.

Ways of reading

Conformance

Identifies whether the digital publication claims to meet internationally recognized conformance standards for accessibility.

Conformance metadata often uses terminology that most people will not understand, and therefore simple statements should be provided when EPUB accessibility and WCAG levels are identified.

If the publication does not include a conformance claim, the statement should indicate that no information is available.

In most cases, people will want to know more about the conformance and certification of the publication. The certifying organization should be identified along with their credentials and placed immediately after the conformance statement.

Navigation

Identifies the navigation features included in the publication.

Navigating a digital publication through a table of contents, an index, by headings, or by go to page shortcuts is important to allow the user to easily access each part of the digital publication. These features are both essential for accessibility and important for all users reading publications. Therefore, clearly communicating to the user what navigation features are available in the publication is important to enable proper title selection.

Note that the Web Content Accessibility Guidelines (WCAG) 2.1 standard requires that more than one way is available to locate a Web page within a set of Web pages (Success Criterion 2.4.5 Multiple Ways).

Rich content

Indicates the presence of math, chemical formulas, extended descriptions for information rich images, e.g., charts, diagrams, figures, graphs, and whether these are in an accessible format or available in an alternative form, e.g., whether math and chemical formulas are navigable with assistive technologies, or whether extended descriptions are available for information-rich images. In addition, it indicates the presence of videos and if closed captions, open captions, or transcripts for prerecorded audio are available.

This group should be displayed only if the metadata indicates the presence of math, chemical formulas, extended descriptions, videos, or transcripts of audio within the title, otherwise it can be hidden.

Hazards

Identifies any potential hazards (e.g., flashing elements, sounds, and motion simulation) that could afflict physiologically sensitive users.

Unlike other accessibility properties, the presence of hazards can be expressed either positively or negatively. This is because users search for content that is safe for them as well as want to know when content is potentially dangerous to them.

The hazards property vocabulary includes a value of unknown, which means the content creator of the metadata explicitly acknowledges that the resource has not been checked for hazards. This is different than providing no metadata for this property.

Accessibility summary

The accessibility summary was intended (in EPUB Accessibility 1.0) to describe in human-readable prose the accessibility features present in the publication as well as any shortcomings. Starting with EPUB Accessibility version 1.1 the accessibility summary became a human-readable summary of the accessibility that complements, but does not duplicate, the other discoverability metadata.

It is a free-form field that allows authors to add additional information to the accessible properties of the resource.

Due to its nature, no specific processing of the content is required; it is sufficient to extract the text from the metadata and display it to end users.

Legal considerations

In some jurisdictions publishers may be able to claim an exemption from the provision of accessible publications, including the provision of accessibility metadata. This should always be subject to clarification by legal counsel for each jurisdiction. One example of this, at time of writing, is from the European Accessibility Act (EAA) when the publisher is a micro-enterprise (i.e., enterprises employing fewer than 10 people and with annual turnover or balance sheet total not exceeding 2 million euro).

Other legal considerations currently included in the EAA may include exceptions on individual versions of a title, where jurisdiction will not require the title to be accessible if it involves a fundamental alteration of the content, or if making it accessible would place a disproportionate burden on the publisher. (This may vary in other jurisdictions).

Publishers may need to include information about an exemption or exception in metadata for legal or clarity reasons, either to bodies that enforce legislation or for other business to business communication. However this is not information that needs to be displayed on public sites as it does not mean anything to most consumers and could lead to misunderstandings. Nevertheless, the objective is to provide as much accessibility information as possible.

Additional accessibility information

This section lists additional metadata categories that can help users better understand the accessibility characteristics of digital publications. These are for metadata that do not fit into the other categories or are rarely used in trade publishing.

Additional accessibility information includes a wide range of information related to the publication's content. Therefore, the features are grouped below so that the presentation is more understandable. Implementors may choose to group the additional information by these categories in their presentations.

The categories are:

Structure

For information on any structuring aids that facilitate use of a resource (e.g., ARIA).

Page breaks

The inclusion of page break markers from a print source allows users to identify where they are in a digital publication relative to its print equivalent.

Adaptation

For information on provisions in the content that enable reading in alternative access modes (e.g., ruby annotations, sign language, transcripts).

Clarity

For information on ways that the content has been enhanced for improved auditory or visual clarity (e.g., pronunciation markup).

Tactile

For information on content that is available in a tactile form (e.g., tactile content, tactile graphics, tactile objects).

Other

For information that does not fall into one of the preceding categories (e.g., timing control).

Minimum filtering set

Reading systems, commerce, and distribution platforms will typically have specific filtering options; having uniformity for key aspects and providing guidance for a standardized approach can help the discovery process for users searching for relevant titles. Accomplishing this, however, should not prevent users with specific reading needs from finding books they are looking for. To achieve this goal, it is recommended that all platforms present two minimum capabilities, centered around the ways of consuming the content. These are:

1. Titles that support visual adjustments
2. Titles that support nonvisual reading

Of note, only the positive values should be used.

Extended filtering set

In specific domains, the addition of other options will be important to help users find content that responds to a particular need or scenario. Each domain case would drive the selection of appropriate items. Some examples of these domains (not exclusive) are:

• In an academic context, the presence of printed page numbers may be crucial to make sure the reader will be able to find or make a citation.
• In a technical or scientific context, information about access rich content like charts, diagrams, figures, graphs, math, and chemical formulas will be important.
• In markets where compliance to legislation is required, understanding conformance information may be mandatory.
• Platforms dedicated to selling children's books and interactive content may want to allow users to select books with no hazards or inform them about the presence of prerecorded audio.

Localization strategy

With the rapidly evolving landscape of accessible ebooks available, most concerned persons are discovering a new world. To ensure the information is widely spread and understood, the implementation by local actors like libraries serving persons with disabilities and other local actors in the fight to end the book famine for readers with print disabilities is crucial and should be highlighted.

The localization of the display guide is a good opportunity to make known the rich accessibility features offered by modern formats like EPUB. In addition, having a local project scale to write an understandable vocabulary to describe such features is a precious qualitative approach that will benefit every player in the value chain.

To build a local project and facilitate feedback comparisons, you can use open source methodology, resources, and prototypes published by EDRLab for the original Signposting accessible digital books project that led to the initial Feedback letter (PDF) addressed in early 2022, and has already been used and adapted in various places.

In these projects, emphasis was placed on the implications for the end users. For instance, the French wording proposed by EDRLab resulted from a quantitative survey of different reader groups, carefully selected panels of individual observations, and an extended feedback process through a dedicated formula available on the 140 first implementation platforms.

Additionally, companies like VitalSource that need wide localization have offered to open-source their professional translation work, which has produced the needed quantity and that is better to use than nothing when no national project has issued a handcrafted vocabulary.

To reconcile both sources of provided localization materials, we propose a collection mechanism based on a detailed identification of the provenance. Because persons and organizations with different levels of technicality must be able to contribute, we accept both raw files via Pull requests and also propose a friendly localization user interface through GitLocalize.

How to contribute?

First let us know as soon as possible that you are working on a localization and wish to submit it. That allows us to prepare a placeholder for your work. This is not mandatory but we invite you to contact the group and participate to a regular call of the working group as those are open to anyone.

When you are ready to publish your work, two options are possible:

• If you don’t know what a JSON or a Pull Request is, you are welcome to contact us so we can attribute a translator role at the Gitlocalize dedicated project page.
• If you feel technically ready or have a collaborator that can push a pull request, the process is to duplicate the canonical original file UX-Guide-Metadata/draft/localizations/en-US/display_guide_vocabulary_w3c_en-US.json, modify it by changing the values in front of each key, and open a pull request so we can review it. Please be aware that we could have questions or ask for precision in the process before accepting and merging your contribution.

How to choose between localization files?

The first keys of each JSON files contains descriptive information about it, including:

• Author, name of the organisation responsible for the establishment and maintenance of this localization
• Language, is a 4 letters code where the two first letters specifies the language as per ISO 639-1 and the two others the country as per ISO 3166-1 alpha-2
• Variant, a one word free name to identify your work
• Audience describes the public. We recommend to use any vocabulary from the ONIX list 28. More than one audience can be informed with a comma separating each.
• Description, a free field including a short description of how this localization was obtained