EPub

Definition

ePub is a popular open e-book standard.

“".epub" is the file extension of an XML format for reflowable digital books and publications. ".epub" is composed of three open standards, the Open Publication Structure (OPS), Open Packaging Format (OPF) and Open Container Format (OCF), produced by the IDPF.” (dipf, retrieved 22:38, 26 February 2009 (UTC))

See also:

• e-book (overview article)
• e-book reader (short hardware overview)
• e-book conversion with Calibre

Software and Formats

ePub can be authored and read with an increasing set of software. Since it is an open standard, it does have support from various vendors and publishers (see e.g. Tim O'Reilly Unplugged: The Kindle 2 And Transforming Industries).

Formats

ePub formats are defined with Relax NG but rely on other standards too.

Overview

The ePub Specification comes in three parts:

(1) The open publication structure (OPS) - 09/11/07

• The Open Publication Structure 2.0 (OPS) is an XML-based standard for authoring digital publications. Contents can be marked up either with a subset of XHTML or Daisy DTBook.
• The Open Packaging Format (OPF) 2.0 describes the structure of an .epub in XML

(2) The open container format - 10/27/06

• The OPS Container Format 1.0 (OCF) is a zip-based standard used to encapsulate publication components for transport and delivery.

ePub contents may be DRM controlled, but must not ...

The container and packaging

The *.epub zip file by example:

If we create an e-pub version of this page we get a file called xxx.epub. This *.epub file is an OCF zip file. Here is the structure:

edutechwiki_epub.epub (the zip file)

META-INF folder

container.xml

OPS folder

fonts folder (includes ttf fonts used)

content1.xhtml

content.opf

style.css

image_name1

image_name2

....

Mimetype

That kind of packaging structure follows quite a similar philosophy as the IMS Content Packaging standard. I.e. a zip file includes a central xml file (content.opf) that includes the definition of organization (the "spine") and the metadata. It then includes all the assets needed for rendering.

The mime-type, i.e. contents of the Mimetype file is application/epub+zip.

Let us now describe some of its files:

The Open Packaging Format (OPF)

File content.opf describes and organizes the various content elements of the epub package. It also provides metadata about the publication, fallback mechanisms when unsupported extensions are used, and a table of contents.

E.g. an example made with an automatic online converter for this page looks like this:

<?xml version='1.0' encoding='UTF-8'?>
<package xmlns:dc="http://purl.org/dc/elements/1.1/"
         xmlns="http://www.idpf.org/2007/opf" version="2.0" unique-identifier="bookid">
  <metadata>
    <dc:title>EPub</dc:title>
    <dc:identifier id="bookid">web2fb2_200904221954_3347363837</dc:identifier>
    <dc:language>En</dc:language>
    <dc:creator>Daniel K. Schneider</dc:creator>
    <dc:type>reference</dc:type>
  </metadata>
  <manifest>
    <item id="css" href="style.css" media-type="text/css"/>
    <item id="content1" href="content1.xhtml" media-type="application/xhtml+xml"/>
    <item id="i0ced13f269" href="i0ced13f269" media-type="image/png"/>
    <item id="ib166f0f69c" href="ib166f0f69c" media-type="image/png"/>
    <item id="i7fa52f212a" href="i7fa52f212a" media-type="image/png"/>
    <item id="i33954a4ae2" href="i33954a4ae2" media-type="image/png"/>
    <item id="i8be224f209" href="i8be224f209" media-type="image/png"/>
  </manifest>
  <spine>
    <itemref idref="content1"/>
  </spine>
</package>

(1) The manifest

This manifest (like in IMS content packaging) must include all files that are part of the publication in any order. It must have a structure like this according to the specification. I.e. each item must have an id, an href to a resource and a media-type. In addition, one can define fall-back elements.

 <manifest>
        <item id="intro" href="introduction.html"
                media-type="application/xhtml+xml" />
        <item id="c1" href="chapter-1.html"
                media-type="application/xhtml+xml" />
        <item id="c2" href="chapter-2.html"
                media-type=application/xhtml+xml" />
        <item id="toc" href="contents.xml"
                media-type="application/xhtml+xml"
                fallback="fall1"  />
        <item id="oview" href="arch.png"
                media-type="image/png" />
        <item id="fall1" fallback="fall2"
                href="SomeDoc.pdf"
                media-type="application/pdf" />
 </manifest>

(2) The spine section

“Following manifest, there must be one and only one spine element, which contains one or more itemref elements. Each itemref references an OPS Content Document designated in the manifest. The order of the itemref elements organizes the associated OPS Content Documents into the linear reading order of the publication.” (Open Packaging Format (OPF) 2.0 specification). This spine (one could translate this to "parts") can include three different kinds of files:

• XHTML
• XML (yours, i.e. what they call out-of-line XML
• DTBook

Spine elements refer to resources defined in the manifest and may include a table of contents. E.g. a simple example would look like this:

<manifest>
     <item id="intro"
           href="intro.html"
           media-type="application/xhtml+xml" />
     <item id="chap1"
           href="chap1.html"
           media-type="application/xhtml+xml" />
     <item id="chap2"
           href="chap2.dtb"
           media-type="application/x-dtbook+xml" />
     <item id="chap3"
           href="chap3.html"
           media-type="application/xhtml+xml" />
     <item id="f1"
           href="fig1.jpg"
           media-type="image/jpeg" />

     <!--  ...... other multimedia assets here .... -->

     <item id="toc_item"
           href="toc.ncx"
           media-type="application/x-dtbncx+xml" />
</manifest>

<spine toc="toc_item">
     <itemref idref="intro" />
     <itemref idref="chap1" />
     <itemref idref="chap2" />
     <itemref idref="chap3" />
</spine>

(3) The metadata section

The metadata are defined using Dublin Core plus possible user-defined tags. Some of these metadata are mandatory, i.e. title, identifier and language.

Other remarks

The XHTML files can include various formats, e.g. binary pictures, SVG and in-line XML. All these formats can be style with a subset of CSS2.

This Open Packaging Format (OPF) 2.0 v0.9871.0 is defined as a relax NG schema.

META-INF files in the epub package

All valid OCF Containers must include a directory called META-INF at the root level of the container file system. This directory contains the files specified below that describe the contents, metadata, signatures, encryption, rights and other information about the contained publication. (OCF 1.0 specification, retrieved 19:17, 22 April 2009 (UTC)).

The container.xml file describes in a simple case where to find the content.opf file. In our simple example it looks like this:

<?xml version="1.0"?>
<container version="1.0" xmlns="urn:oasis:names:tc:opendocument:xmlns:container">
  <rootfiles>
    <rootfile full-path="OEBPS/content.opf" media-type="application/oebps-package+xml"/>
  </rootfiles>
</container>

An other example taken from the OCF 1.0 specification show that one could include an alternative PDF file for example:

<?xml version="1.0"?>
<container version="1.0" xmlns="urn:oasis:names:tc:opendocument:xmlns:container">
  <rootfiles>
    <rootfile full-path="OEBPS/My Crazy Life.opf"
     media-type="application/oebps-package+xml" />
    <rootfile full-path="PDF/My Crazy Life.pdf"
     media-type="application/pdf" />
  </rootfiles>
</container>

In addition to container.xml, there can be five other files:

• manifest.xml
• metadata.xml
• signatures.xml
• encryption.xml
• rights.xml

The formal specification of these files in META-INF is done with a little Relax NG schema:

• There is one rule for container (see the examples above).
• One rule for signatures and that includes the xmldsig-core-schema.rng Schema
• One rule for encryption that also refers to extenal *.rng Schemas.

The XHTML modules used in OPS

OPS uses a set of XHTML modules with some additional restrictions. E.g. OPS is always XHTML compatible, but no the other way round.

XHTML 1.1 Module Name	Elements (non-normative)
Structure	body, head, html, title
Text	abbr, acronym, address, blockquote, br, cite, code, dfn, div, em, h1, h2, h3, h4, h5, h6, kbd, p, pre, q, samp, span, strong, var
Hypertext	a
List	dl, dt, dd, ol, ul, li
Object	object, param
Presentation	b, big, hr, i, small, sub, sup, tt
Edit	del, ins
Bidirectional Text	bdo
Table	caption, col, colgroup, table, tbody, td, tfoot, th, thead, tr
Image	img
Client-Side Image Map	area, map
Meta-Information	meta
Style Sheet	style
Style Attribute (deprecated)	style attribute
Link	link
Base	base

Remark: EPub also can use the DTBook (DAISY/NISO standard) for markup.

SVG

Readers must support SVG 1.1. SVG animation and scripting features are not supported and must not be used by publication authors; a Reading System should not render such content. CSS styling of SVG must be fully supported.

SVG content can be used from XHML img and object elements but also within XHTML (probably in the standard way with namespaces).

XML

You also may use your own XML both inline within XHTML and out-of-line as documents. Both can have fallback options (to be used when the contents can't be rendered by a client).

OPS style Sheets

OPS Style sheets are CSS2 styles in the XML tradition, i.e. selectors and attribute names are case sensitive. Again, like for XHTML, there are some restrictions.

Software

Authoring software

Sigil

Sigil seems to be the best free authoring software so far.

“Web-standard formats such as XHTML make up the core of OPS. OPS files can be created using a wide variety of Web and eBook publishing tools. They can also be created using XML editors such as Altova’s XML Spy. Adobe InDesign CS3 supports the direct generation of OCF-packaged OPS content as an export function. OCF files can also easily be created using standard ZIP applications like WINZIP” (IDPF FAQ, retrieved 22:38, 26 February 2009 (UTC).

Probably the best solution for authoring new e-pub contents is using XHTML authoring tool and then use conversion software. One of the problems is that one should produce several smaller XHTML files since epub is meant for small devices. Another is managing graphic assets like pictures.

To create the epub archive, there are two options so far:

• Build the epub archive manually which is not too difficult for a technical person.
• Use conversion software (see below), some of which is good. E.g. Calibre can handle multi-file documents in various input formats and it also can split a well structured XHTML file with an XPath expression into several sub-files.

eScape (Not tested so far)

• EScape is a free for non-commercial use Open Office Writer to ePub creator. According to the -ODT2ePub/ home page, production is easy:

1. Use the supplied Open Office Template (OTT) to Structure-Style™= your text files in the familiar Open Office Writer environment.
2. Open eScape, browse to your ODT, cover image and select a CSS.
3. Set a few reader friendly options and click "Create ePub"

See also: Atlantis Word processor (below)

Conversion software

There exist several tools that can convert from one format to another. Mileage varies according to input. E.g. PDF is more difficult to convert than XHTML for example,

• PDF input leads to really ugly output
• Web pages like this mediawiki page translate badly, e.g. bullets go away.

Atlantis Word Processor

Atlantis Word Processor converts any document to EPUB. Supports multilevel TOCs, font embedding, and batch conversion.

Source formats: RTF, DOC, DOCX, ODT, TXT.

Platform: Windows.

Calibre

Calibre is a one stop solution to all your e-book needs. It is free, open source and cross-platform in design and works well on Linux, OS X and Windows.

List of general features: Library Management - Format conversion (all major ebook formats can be converted from) - Syncing to ebook reader devices - Fetching news from the web and converting it into ebook form - Viewing many different ebook formats - Giving you access to your book collection over the internet using just a browser.

Converter: can create chapters out of single text using XPath expressions, add title image, title, author, data, etc.

Source formats: LIT, MOBI, EPUB, HTML, PRC, RTF, TXT, PDF. Some convert better than others, e.g. PDF pictures don't translate.

Output formats: EPUB, LRF, MOBI

Easy to install under Ubuntu (tested with 8.04 Hardy Heron and 9.04 Jaunty). You can do it with one command line.

• Read e-book conversion with Calibre for a short how-to.

eCub

eCub a simple to use EPUB and MobiPocket ebook creator.

Cross-platform: Win/Mac/Linux/etc.

Converter can make a book out of several HTML files and add Author, Title picture, etc.

Input formats: txt and XHTML

Output formats: EPUB and MobiPocket

Extra features: Can convert book content to audio MP3 and WAV.

Easy to install under Ubuntu (tested with 8.04 Hardy Heron and 9.04 Jaunty). All you need to do is to click on the debian distribution link and install.

On-line conversion tools

• Web2FB2 is a webpage to F3B and EPUB converter. (tested April 2009).
  • The result wasn't convincing regarding conversion of a fairly ugly XHTML file (I tried with this page) - Daniel K. Schneider 19:17, 22 April 2009 (UTC).
  • Tip: make use of the advanced options to set both title and author. The file name created will be named authorname_first_name_title.epub.

Wiki 2 epub

• We did not find any software, but as we describe below, wiki to xhtml to epub somewhat works.

Creation and conversion testing

A (preliminary) conversion test of a huge HTML file

(Daniel K. Schneider 19:17, 22 April 2009 (UTC))

I tried to convert mediawiki contents, i.e. what could be called a wiki book.

• I took a larger list of flash tutorials from this wiki (bad idea since they include huge pictures) and generated a single (ugly) html page with the pdfbook generator (800MB for the HTML only).
• This some_flash_tutorials.html file then had to be repaired with tidy. I also converted it to real XHTML.

tidy -o flash_tutorials.xhtml -asxhtml some_flash_tutorials.html

• I opened the XHTML file in a browser and then "saved as complete web page" in a directory. This is important in order to have a local copy of all images.
• I then imported the XHTML File to Calibri and converted to epub.
  • I changed the XPath expression for chapter detection to "//h1" since each each wiki page starts with h1.

The result was sort of ok, a "442 pages" 8MB file

• Some links ("a" tags) were wrong, i.e. extented over several pages. They were ok in the original XHTML file. The conversion software cannot handle something like this and looses a closing "a" somewhere.

 <div class="thumb tleft">
  <div class="thumbinner" style="width: 182px;">
  <a href="http://edutechwiki.unige.ch/en/Image:Flash-cs3-tools-panel-items.png"
     class="image" title="Items of the Flash CS3 tools panel">
  <img height="480" border="0" width="180" 
       alt="Items of the Flash CS3 tools panel"
       src="flash_tutorials_files/180px-Flash-cs3-tools-panel-items.png"
       class="thumbimage"/></a>
  <div class="thumbcaption">
  <div class="magnify">
  <a href="http://edutechwiki.unige.ch/en/Image:Flash-cs3-tools-panel-items.png"
     class="internal" title="Enlarge"/></div>
 Items of the Flash CS3 tools panel</div>
 </div>
</div>

• There were too many wiki links (normal since my wiki pages are linked).

Basically, I'd have to clean up the XHTML to get a better result, i.e. remove some wiki things that are really not needed.

I then also tested eCub.

• Copy the xhtml file plus the image directory into the project directory.
• In Options, tick Portable mode. This will include images

The result was a also a 8.7MB file. This software is a bit easier to use and results were slightly better, but the links problem was the same. But it can't split a file into chapters, i.e. create a table of contents for a single big HTML file. Therefore, one has to import xhtml files one by one in order to get a chapter structure.

Conclusion: It is possible to create rather large e-books from mediawiki pages. But for quality results there is manual work to be done (or filtering script writing).

Reader Software

Most readers support several e-book formats and several support ePub. See also: EPUB at mobileread.com and Wikipedia's Comparison of e-book formats.

There exist several ePub capable readers. Here are just some of these:

FBReader

FBReader e-book reader for Unix/Windows computers. It supports several formats.

Wikipedia entry

Unix/Windows (It's included in Ubuntu's Synaptic)

Features: Several formats (partially or fully) . Several reading options.

Calibre

Multi-purpose tool, see above

Openberg

OpenBerg, was an opensource initiative to write a system (reader, authoring etc.). The reader was a Firefox extension. Now dead ?

Wikipedia entry

Adobe Digital editions

Adobe Digital Editions is a free software with built-in DRM mechanism. If you buy contents and if you register with Adobe you can use the contents on a limited set of other computers - needs some clarification)

Wikipedia entry

Stanza

Lexcycle Stanza (favorite free iPhone/iPod reader). Win/Mac/i*.

Wikipedia Entry

Hardware

E-books make most sense when read on specialized hardware. Several brands can read ePub documents, e.g.

Sony Reader (Mobileread wiki)

E.g. the Sony PRS 505 has 64MB RAM and 256 flash RAM, a size of 6", 800 x 600px, 167 PPI, 8 grey levels, USB 2.0, about 2-3 weeks battery life (7500 page turns).

E.g. the Sony PRS 700BC (oct 2008) has a size of 127.6 x 174.3 x 9.7 mm, a touch screen / 167 ppi / 800x600px.

JetBook

153 x 109 x 10mm: 640 x 480 VGA greyscale

Hanlin EBooks (Mobileread wiki).

The same hardware is sold under different brand names and with different firmware.

See e-book.

Links

General

• EPUB (mobileread.com wiki)

Official

• International Digital Publishing Forum (idpf, where the standards are defined)

Software

• epubtest has a good list of software, including commercial and free authoring/conversion tools

Contents

• Where to download ePub Books online. Several publishers are listed, but also free contents.
• Feedbooks (free contents)

Tutorials and technical tips

• The ABCs of e-book format conversion: Easy Calibre tips for the Kindle, Sony and Nook By John Schember