The section element is used to encapsulate sections of primary content and establish the
publication's structural hierarchy. Clearly grouping and identifying content enables better navigation by
anyone reading using an assistive technology (see also the faq).
An epub:type attribute should be attached to each section to indicate the
specific nature of the content, when applicable. See the EPUB 3 Structural Semantics Vocabulary for the list of allowed values that can be used with the
attribute.
When the epub:type attribute is not specified, a generic section/subsection nature is assumed
based on the nesting level.
A section must not have more than one child heading.
If a section does not have a heading, consider including an aria-label attribute on the
section to identify its purpose.
<section epub:type="chapter" id="c01">
<h1 id="c01h01">Chapter 1. Loomings.</h1>
<p>Call me Ishmael. … </p>
…
</section>title attribute
<section epub:type="chapter" title="chapter" id="c01">
<p>It was a dark and stormy night … </p>
…
</section>aria-label attribute
<section epub:type="chapter" aria-label="chapter" id="c01">
<p>It was a dark and stormy night … </p>
…
</section>section element for secondary content?No, the section element should only be used to identify primary structural
content.
Self-contained secondary content embedded within a section — such as sidebars,
mini-articles, etc. — should be identified using the aside and
article elements, as appropriate.
EPUBs are typically broken up into smaller XHTML content files to improve rendering, so it's often the case that an entire part cannot be contained in a single file. The question becomes, how do you split the chapters into separate files and still retain the part?
There are two approaches you can take to flattening out your content, and neither is optimal — at least in terms of retaining the original structural hierarchy in the markup. The first is to include the part heading with the first chapter:
<body>
<section epub:type="part">
<h1>Part I</h1>
<section epub:type="chapter">
<h2>Chapter I</h2>
</section
</section>
</body>Each subsequent chapter file would contain a single section with an h2
heading. Although this markup approach appears to suggest that the part only contains a single
chapter, consistent heading numbering will allow an assistive technology to correctly orient the
reader.
The other approach is to separate the part heading into its own file:
<body>
<section epub:type="part">
<h1>Part I</h1>
</section>
</body>Again, each subsequent chapter file would include an h2 heading to indicate it is
subordinate to the part. One advantage of this approach is that reading systems will render the
part heading separately from the first chapter content.
It would be nice to see a future version of EPUB include a way of retaining information about the document hierarchy regardless of how the content gets broken up into files (e.g., in the spine), but at this time no such mechanism exists.
Yes and no.
Grouping content into section elements more clearly delineates the structure of your
document, since there's no ambiguity about where the section begins and ends (e.g., in HTML4,
headings were often interspersed throughout the content regardless of relation to structure).
This structure allows readers using assistive technologies to move from section to section more
easily.
It's also true that grouping and identifying the nature of each section with an
epub:type attribute aids navigation of the document at the markup level (which
is where assistive technologies usually operate).
That said, because the publication's hierarchy often gets flattened when content is split into
separate XHTML files, the section element alone is left somewhat crippled as a
navigation aid. A reader will have more difficulty moving quickly from part to part and skipping
all the chapters in between unless they use the table of contents (i.e., each content file would
have to be loaded and each section type checked to move forward at the markup
level).
See the Headings section for more information.