EPUB 3 Overview

› 1.1 Overview

The EPUB® specification is a distribution and interchange format standard for digital publications and documents. EPUB defines a means of representing, packaging and encoding structured and semantically enhanced Web content — including HTML5, CSS, SVG, images, and other resources — for distribution in a single-file format.

EPUB 3, the third major release of the standard, consists of a set of four specifications, each defining an important component of an overall EPUB Publication:

EPUB has been widely adopted as the format for digital books (eBooks), and these new specifications significantly increase the format's capabilities in order to better support a wider range of publication requirements, including complex layouts, rich media and interactivity, and global typography features. The expectation is that EPUB 3 will be utilized for a broad range of content, including books, magazines and educational, professional and scientific publications.

This document provides a starting point for content authors and software developers wishing to understand these specifications. It consists of non-normative overview material, including a roadmap to the four building-block specification documents that compose EPUB 3.

Another non-normative document, EPUB 3 Changes from EPUB 2.0.1 [EPUB3Changes], describes changes in EPUB 3 from the previous version, but is intended primarily for Authors and EPUB Reading System vendors migrating from EPUB 2.0.1 to EPUB 3 and for those who anticipate supporting both versions.

› 1.2 Roadmap

This section provides an overview of the EPUB 3 specifications by explaining in brief the components of a Publication. Links to additional information within this document and to the specifications are included.

An EPUB Publication, at its most basic level, is a bundled collection of resources that can be reliably and predictably ingested by an EPUB Reading System in order to render its contents to a User. Some of these resources facilitate the discovery and processing of the Publication, while others make up the content of the publication. The latter, EPUB Content Documents, are described in Content Documents and are fully defined in [ContentDocs30].

A Publication's resources are typically bundled for distribution as a ZIP-based archive with the file extension .epub. As conformant ZIP archives, Publications can be unzipped by many software programs, simplifying both their production and consumption. The container format is introduced in Container and defined in [OCF3].

The container format not only provides a means of determining that the zipped content represents an EPUB Publication (the mimetype file), but also provides a universally-named directory of informative resources (/META-INF). Key among these is the container.xml file, which directs Reading Systems to the root file of the Publication, the Package Document.

The Package Document is itself a kind of information warehouse for the Publication, storing metadata about the specific work contained in the Publication, providing an exhaustive list of resources and defining a default reading order. The Package Document is introduced in Package Document and defined in [Publications30].

The preceding components of an EPUB Publication are not new to EPUB 3, and will be familiar to anyone who has worked with Publications before, although they have been changed and enhanced in this version. A new core addition to EPUB 3, however, is the Media Overlay Document, which defines a means of synchronizing text and audio playback. The Overlay Document is introduced in Aural Renditions and Media Overlays and defined in [MediaOverlays30].

The following example shows the resources a minimal "Hello World" Publication might contain:

mimetype
META-INF/container.xml
Content/HelloWorld.opf
Content/HelloWorld.xhtml

While conceptually simple, an EPUB Publication is more than just a collection of HTML pages and dependent assets in a ZIP package as represented in this example. The following sections of this document delve into more detail about the primary features and functionality that Publications provide to enhance the reading experience.

› 2.1 Package Document

Every EPUB Publication includes a single Package Document, which specifies all the Publication's constituent content documents and their required resources, defines a reading order for linear consumption, and associates publication-level metadata and navigation information.

The Package Document represents a significant improvement on a typical Web site. A Web site, for example, embeds references to its resources within its content, which, while a simple and flexible means of identifying resources, makes it difficult to enumerate all the resources required to render it. In addition, there is no standard way for a Web site to define that a sequence of pages make up a larger publication, which is precisely what EPUB's spine element does (i.e., it provides an external declarative means to explicitly specify navigation through a collection of documents). Finally, the Package Document defines a standard way to represent metadata globally applicable to a collection of pages.

The Package file and other publication-level constructs are specified in [Publications30].

› 2.2 Navigation

› 2.3 Metadata

EPUB Publications provide a rich array of options for adding Publication metadata. The Package Document includes a dedicated metadata section [Publications30] for general information about the Publication, allowing titles, authors, identifiers and other information about the Publication to be easily accessed. It also provides the means to attach complete bibliographic records to a Publication using the link element [Publications30]).

The Package Document also allows a Unique Identifier to be established for a Publication using the unique-identifier attribute [Publications30]. The required last-modified date in the Package metadata section can be joined with this identifier to define a Package Identifier, which provides a means of distinguishing EPUB Publications that represent different versions of the same Manifestation (see Publication Identifiers [Publications30]). The Package Identifier addresses the issue of how to release a Publication without changing its Unique Identifier while still identifying it as a new version.

XHTML Content Documents also include the means of annotating document markup with rich metadata, making them more semantically meaningful and useful both for processing and accessibility purposes (Semantic Inflection [ContentDocs30]).

› 2.4 Content Documents

Every EPUB Publication contains one or more EPUB Content Documents, as defined in [ContentDocs30]. These are XHTML or SVG documents that describe the readable content of a Publication and reference associated media resources (e.g., images, audio and video clips).

XHTML Content Documents are defined by a profile of HTML5 that requires the use of XML serialization [HTML5] in order to ensure that content can be reliably manipulated and rendered. This profile also adds two additional EPUB-specific language constructs: the epub:type attribute [ContentDocs30] for element-level metadata and the epub:trigger element [ContentDocs30] for declaratively associating controls with multimedia elements.

These additions do not affect the ability of an HTML5 User Agent [HTML5] to render EPUB XHTML Content Documents, but Publications may not render identically in all User Agents depending on their support.

› 2.5 Rendering and CSS

A key concept of EPUB is that content presentation should adapt to the User rather than the User having to adapt to a particular presentation of content. HTML was originally designed to support dynamic rendering of structured content, but over time HTML as supported in Web browsers has become focused on the needs of Web applications, and most popular Web sites now have fixed-format layouts.

EPUB Publications, however, are designed to maximize accessibility for the visually impaired, and Reading Systems typically perform text line layout and pagination on the fly, adapting to the size of the display area, the User's preferred font size, and other environmental factors. This behavior is not guaranteed in EPUB; images, vector graphics, video and other non-reflowable content may be included, and some Reading Systems may not paginate on the fly, or at all. Nevertheless, supporting dynamic adaptive layout and accessibility has been a primary design consideration throughout the evolution of the EPUB standard.

EPUB Content Documents may optionally reference EPUB Style Sheets, allowing Authors to define the desired rendering properties. EPUB 3 defines a profile of CSS based on CSS 2.1 [CSS2.1] for this purpose, together with capabilities defined by various CSS3 Modules and several additional properties specific to EPUB.

CSS3 properties were selected based on their current level of support in Web browsers, but support for them in Reading Systems and User Agents is not guaranteed (EPUB-defined properties may similarly be ignored).

EPUB 3 also supports CSS styles that enable both horizontal and vertical layout and both left-to-right and right-to-left writing, but Reading Systems may not support all of these capabilities. Reading Systems may also support different rendering options than the Author intended. Refer to CSS in the Global Language Support section for more information.

EPUB 3 also supports the ability to include multiple style sheets that allow users, for example, to select between day/night reading modes or to change the rendering direction of the text. Refer to Alternate Style Tags [ContentDocs30] for more information.

› 2.6 Fonts

EPUB 3 supports two closely-related font formats — OpenType [OpenType] and WOFF [WOFF] — to accommodate both traditional publishing workflows and emerging Web-based workflows. Word processing programs used to create Publications are likely to have access only to a collection of installed OpenType fonts, for example, whereas Web-archival EPUB generators will likely only have access to WOFF resources (which cannot be converted to OpenType without undesirable, and potentially unlicensed, stripping of WOFF metadata).

EPUB 3 also supports both obfuscated and regular font resources for both OpenType and WOFF font formats. Support for obfuscated font resources is required to accommodate font licensing restrictions for many commercially-available fonts.

› 2.7 Scripting

EPUB strives to treat content declaratively — as data that can be manipulated, not programs that must be executed — but does support scripting as defined in HTML5 and SVG (refer to Scripted Content Documents [ContentDocs30] for more information).

It is important to note, however, that scripting support is optional for Reading Systems and may be disabled for security reasons.

Authors should also note that scripting in an EPUB publication can create security considerations that are different from scripting within a Web browser. For example, typical same-origin policies are not applicable to content that has been downloaded to a User's local system. Therefore, it is strongly encouraged that scripting be limited to container constrained contexts, as further described in Scripted Content Documents — Content Conformance [ContentDocs30].

Scripting consequently should be used only when essential to the User experience, since it greatly increases the likelihood that content will not be portable across all Reading Systems and creates barriers to accessibility and content reusability.

› 2.8 Text-to-speech

EPUB 3 provides the following text-to-speech (TTS) facilities for controlling aspects of speech synthesis, such as pronunciation, volume, pitch and rate:

› 2.9 Container

An EPUB Publication is transported and interchanged as a single file (a "portable document") that contains the Package Document, all Content Documents and all other required resources for processing the Publication. The single-file container format for EPUB is based on the widely adopted ZIP format. An XML manifest that specifies the location in the ZIP archive of the package file must be found at a well-defined location within the archive.

This approach provides a clear contract between any creator of an EPUB publication and any system which consumes such publications, as well as a reliable representation that is independent of network transport or file system specifics.

An EPUB publication's representation as a container file is specified in [OCF3].

› 3.1 Metadata

EPUB 3 supports alternate representations of all text metadata items in the package metadata section to improve global distribution of Publications. The alternate-script property can be combined with the xml:lang attribute to include and identify alternate script renditions of language-specific metadata.

Using this property, a Japanese publication could, for example, include an alternate Roman-script representation of the author's name and/or one or more representations of the title in Romance languages. Refer to the alternate-script property [Publications30] for more information.

The page-progression-direction attribute also allows the content flow direction to be globally specified for all Content Documents to facilitate rendering (see the page-progression-direction [Publications30]).

› 3.2 Content Documents

XHTML Content Documents leverage the new HTML5 directionality features to improve support for bidirectional content rendering: the bdi element allows an instance of directional text to be isolated from the surrounding content, the bdo element allows directionality to be overridden for its child content and the dir attribute allows the directionality of any element to be explicitly set.

XHTML Content Documents also support ruby annotations for pronunciation support (which makes them supported in Navigation Document links, as well).

SVG Content Documents support the rendering of bidirectional text, but do not include support for ruby.

› 3.3 CSS

EPUB 3's support for new CSS3 modules enables typography for many different languages and cultures. Some specific enhancements include:

› 3.4 Fonts

EPUB 3 does not require that Reading Systems come with any particular set of built-in system fonts. As occurs in Web contexts, Users in a particular locale may have installed fonts that omit characters required for other locales, and Reading Systems may utilize intrinsic fonts or font engines that do not utilize operating system installed fonts. As a result, the text content of a Publication may not natively render as intended on all Reading Systems.

To address this problem, EPUB 3 supports the embedding of fonts to facilitate the rendering of text content, and this practice is recommended in order to ensure content is rendered as intended.

Support for embedded fonts also ensures that Publication-specific characters and glyphs can be embedded for proper display.

› 3.5 Text-to-speech

EPUB 3's support for PLS documents and SSML attributes increases the pronunciation control that Authors have over the rendering of any natural language in text-to-speech-enabled Reading Systems. Refer to Text-to-speech in the Features section for more information on these capabilities.

The combination of CSS Speech and inline SSML phonemes also allows fine control over ruby.

› 3.6 Container

The OCF container format supports UTF-8, allowing for internationalized file and directory naming of content resources.

› 4.1 Navigation

EPUB 3 improves on NCX documents with the addition of EPUB Navigation Documents. As noted in Content Documents above, the new features these provide represent a more universal and flexible navigation system.

The need to navigate a document is not exclusively an accessibility issue, but a complete Navigation Document is key for all Users to be able to easily and fully access the contents of a document. The primary toc nav element (as defined in The toc nav Element [ContentDocs30]) should fully reflect the structure of the Publication, as a result.

For highly structured documents where it may not be desirable to display the complete table of contents to Users, the display level can be modified using the CSS display property without minimizing the information that is available for accessible Reading Systems.

Authors are also encouraged to supply additional nav elements if their Publications contain non-structural points of interest, such as figures, tables, etc. in order to further enhance access to the content.

› 4.2 Semantic Markup

HTML5 supports a number of new elements intended to make markup more semantically meaningful (e.g., section, nav, aside) and introduces more clearly defined semantics for some HTML4 elements. These elements, in conjunction with best practices for authoring well-structured Web content, should be utilized when creating EPUB XHTML Content Documents. These additions allow content to be better grouped and defined, both for representing the structure of documents and to facilitate their logical navigation. XHTML Content Documents also natively support the inclusion of ARIA role and state attributes and events, enhancing the ability of Assistive Technologies to interact with the content.

EPUB 3 further introduces the epub:type [ContentDocs30] attribute, which is meant to be functionally equivalent to the W3C Role Attribute [Role]. This attribute allows any element in an XHTML Content Document to include additional information about its purpose and meaning within the work, using controlled vocabularies and terms. Refer to Semantic Inflection [ContentDocs30] for more information.

› 4.3 Dynamic Layouts

The design center of EPUB is dynamic layout: content is typically intended to be formatted on the fly rather than being typeset in a paginated manner in advance (i.e., expecting a particular sized "page"). This core capability is useful, for example, for optimizing rendering onto different sized device screens or window sizes, and it facilitates and simplifies content accessibility.

While it is possible to incorporate more highly formatted content in EPUB — for example via bitmap images or SVG graphics, or even use of CSS explicit positioning and/or table elements to achieve particular visual layouts — Authors are strongly discouraged from utilizing such techniques. They are not reliable in EPUB since many Reading Systems render content in a paginated manner rather than creating a single scrolling Viewport and since each Reading System may define its own pagination algorithm. If these techniques are required to convey the content of the publication (for example, for graphic novels), fallbacks [Publications30] should always be included.

In general, it is preferable to achieve visual richness by using EPUB Style Sheets without absolute sizing or positioning.

› 4.4 Aural Renditions and Media Overlays

Aural renditions of content are important for accessibility and are a desirable feature for many other Users. A baseline to facilitate aural rendering is to utilize semantic HTML designed for dynamic layout. Refer to Text-to-speech for more information on how to use the native facilities that EPUB XHTML Documents include.

One of the key new accessibility features in EPUB 3 is the inclusion of Media Overlay Documents [MediaOverlays30]. When pre-recorded narration is available for a Publication, Media Overlays provide the ability to synchronize pre-recorded audio narration with the text of a Content Document, a feature already familiar to readers of Digital Talking Books.

› 4.5 Fallbacks

Not all formats are accessible in their native format, and not all Users prefer to read in the default format provided. EPUB defines a variety of means for providing fallbacks so that alternate renditions of a Publication can be made available in these cases.

Publication and content-level fallbacks are defined in Restrictions and Fallbacks [Publications30]. These allow for the alternate rendition of specific resources within a Publication, such as SVG images or video clips.

In addition, multiple instances of a complete work can be delivered in a single Publication by defining multiple rootfile elements in the OCF container file (as described in Container – META-INF/container.xml [OCF3]). This kind of fallback may be used, for example, so that a formatted graphic novel defined via a sequence of SVG pages can be accompanied by an accessible text version defined via XHTML.

› 4.6 Scripting

EPUB 3 adopts a progressive enhancement approach for scripted content, whereby scripting must not interfere with the integrity of the document (i.e., must not result in information loss when scripting is not available). Consequently, although documents that do employ scripting may provide fallbacks [ContentDocs30] to further facilitate access to their contents, the documents must be accessible without them.

Several mechanisms in EPUB can further minimize and constrain scripting within Publications to improve accessibility:

Best practices for accessible scripting in Web documents, such as provided in [WAI-ARIA], should always be consulted, and use of scripting should be reserved for situations in which interactivity is critical to the User experience.