Filesystem Layout and File Naming Conventions
A bare Standard Ebooks directory structure looks like this:
XHTML files containing the actual text of the ebook are located in
./src/epub/text/. All files in this directory end in
CSS files used in the ebook are located in
./src/epub/css/. All files in this directory end in
.css. This directory contains only two CSS files:
./src/epub/css/core.cssis distributed with all ebooks and is not edited.
./src/epub/css/local.cssis used for custom CSS local to the particular ebook.
Raw source images used in the ebook, but not distributed with the ebook, are located in
./images/. These images may be, for example, very high resolution that are later converted to lower resolution for distribution, or raw bitmaps that are later converted to SVG for distribution. Every ebook contains the following images in this directory:
./images/titlepage.svgis the editable titlepage file that is later compiled for distribution.
./images/cover.svgis the editable cover file that is later compiled for distribution.
./images/cover.source.(jpg|png|bmp|tif)is the raw cover art file that may be cropped, resized, or otherwise edited to create
./images/cover.jpgis the final edited cover art that will be compiled into
Images compiled or derived from raw source images, that are then distributed with the ebook, are located in
The table of contents is located in
The epub metadata file is located in
The ONIX metadata file is located in
./src/epub/onix.xml. This file is identical for all ebooks.
./src/mimetypedirectory and files are epub structural files that are identical for all ebooks.
./LICENSE.mdcontains the ebook license and is identical for all ebooks.
XHTML file naming conventions
Numbers in filenames don’t include leading
Files containing a short story, essay, or other short work in a larger collection, are named with the URL-safe title of the work, excluding any subtitles.
Work Filename A short story named “The Variable Man”
A short story named “The Sayings of Limpang-Tung (The God of Mirth and of Melodious Minstrels)”
Works that are divided into larger parts (sometimes called “parts,” “books,” “volumes,” “sections,” etc.) have their part divisions contained in individual files named after the type of part, followed by a number starting at
Works that are composed of chapters, short stories, essays, or other short- to medium-length sections have each of those sections in an individual file.
Chapters not contained in separate volumes are named
Nis the chapter number starting at
Section Filename Chapter 1
Chapters contained in separate volumes, where the chapter number re-starts at 1 in each volume, are named
Xis the part number starting at
Nis the chapter number within the part, starting at
Section Filename Part 1
Part 1 Chapter 1
Part 1 Chapter 2
Part 1 Chapter 3
Part 2 Chapter 1
Part 2 Chapter 2
Chapters contained in separate volumes, where the chapter number does not re-start at 1 in each volume, are named
Nis the chapter number, starting at
Section Filename Part 1
Works that are composed of extremely short sections, like a volume of short poems, are in a single file containing all of those short sections. The filename is the URL-safe name of the work.
Section Filename A book of short poems called “North of Boston”
Frontmatter and backmatter sections have filenames that are named after the type of section, regardless of what the actual title of the section is.
Section Filename A preface titled “Note from the author”
If a work contains more than one section of the same type (for example multiple prefaces), the filename is followed by
Nis a number representing the order of the section, starting at
Section Filename The work’s first preface, titled “Preface to the 1850 Edition”
The work’s second preface, titled “Preface to the Charles Dickens Edition”
se lint tool makes best guesses to alert the user to potential issues in an ebook production, and it may sometimes guess wrong. An
se-lint-ignore.xml file can be placed in the ebook root to make
se lint ignore specific error numbers in an ebook.
se-lint-ignore.xmlis optional. If it exists, it is in the ebook root.
se-lint-ignore.xmlfile looks like this:
<se-lint-ignore>root element contains one or more
<file>elements have a
pathattribute containing a filename to match in
pathattributes accept shell-style globbing to match files.
<file>element contains one or more
<ignore>element contains one
<code>element and one
The value of
<code>is the error/warning code provided by
se lint. This code will be ignored for its parent file(s) when
se lintis next run.
The value of
<reason>is a prose explanation about why the code was ignored. This aids future producers or reviewers in understanding the reasoning behind why an error code was ignored.
<reason>is required to have a non-whitespace value.
The following is an example of a complete
se-lint-ignore.xml file from Tractatus Logico-Philosophicus.