The Epub3.0 specification defines both what can go into a Epub document and how it should be rendered. It utilizes standard web technologies like CSS3 and JavaScript but uses the XML-version of HTML5. Work on the standard began in 2010.
It can be used for every type of publication and can be read non-linearly. An Epub can contain several diverent versions/publications.
Since it is based on web standards it allows for reflowable content as opposed to e.g. PDFs fixed format. The standard requires MP3 and MathML support and allows for both WebM and H.264.
The specification is comprised of four parts:
- EPUB Publications 3.0
- Required content types.
- The
XML schemathat defines how and what meta information about the document can be provided. - When and how to provide fallbacks.
- EPUB Content Documents 3.0
- A whitelist of
XHTML5,SVG1.1,CSS3features that can be used ("profiles"). - How the document navigation file must be structured.
- A whitelist of
- EPUB Media Overlays 3.0
- How to synchronize the display area with audio narration.
- EPUB Open Container Format (OCF) 3.0
- How to bundle all files together.
- Exactly one OCF container definition file (
META-INF/container.xml)
- Needs to point the package-document.
- Exactly one Package Document
- Needs to lists all referenced resources
- Needs to hold a certain set of document metadata.
- Needs to define the recommended reading order (spine).
- At least one Epub Content Document
- Exactly one Epub Navigation Document
- Exactly one mimetype file (named
mimetype)- Needs to containing the string:
application/epub+zip
- Needs to containing the string:
- Zip the package
- Needs to include all resources except audio and video files wich can be loaded externally.
<?xml version='1.0' encoding='utf-8'?>
<container xmlns="urn:oasis:names:tc:opendocument:xmlns:container" version="1.0">
<rootfiles>
<rootfile media-type="application/oebps-package+xml"
full-path="<!-- package-document -->" />
<!-- ... -->
<!-- You could have more than one rootfile
since you can have multiple publications
per Epub package -->
</rootfiles>
</container>The order of elements is of significance in the package document:
<metadata><manifest><spine>
<?xml version="1.0" encoding="UTF-8"?>
<package xmlns="http://www.idpf.org/2007/opf" version="3.0"
unique-identifier="pub-id">
<metadata xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:dcterms="http://purl.org/dc/terms/"
<!-- ... -->
</metadata>
<manifest>
<!-- ... -->
</manifest>
<spine>
<!-- ... -->
</spine>
</package><dc:identifier id="id">
<!-- unique identifier freely chosen by you,
you could use your ISBN -->
</dc:identifier>
<dc:title id="titel">
<!-- ... -->
</dc:title>
<dc:language>
<!-- IETF language tag -->
</dc:language>
<meta property="dcterms:modified">
<!-- timestamp of last modification,
helps reader track changes
eg. 2012-06-08T14:44:00Z -->
</meta><manifest>
<item media-type="application/xhtml+xml"
id="<!-- ... -->"
href="<!-- ... -->" />
<!-- ... -->
</manifest><item href="<!-- ... -->"
id="image" fallback="NO-image"
media-type="image/png" />
<item href="<!-- ... -->"
id="NO-image"
media-type="application/xhtml+xml"/>In the fallback property you can reference the id of another item which will be used in case the original item can not.
<spine>
<itemref idref="<!-- ... -->" />
<!-- ... -->
</spine>The property idref references elements from the <manifest> due to their id property.
The ordering of the itemref elements represents the recommended reading order.
Since Epub3 uses XHTML5 file should end with .xhtml and have the following opening. The whole set of HTML5 elements can be used.
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html>
<html xmlns=http://www.w3.org/1999/xhtml>
<!-- ... -->The extensions are defined in the EPUB3 Structural Semantics Vocabulary. If the extensions will be used the <html> element needs to extended with the xmlns:epub property:
<html xmlns="http://www.w3.org/1999/xhtml"
xmlns:epub="http://www.idpf.org/2007/ops">epub:type="chapter"epub:type="titleepub:type="subtitle"epub:type="contributors"epub:type="foreword"