How HTML Changes in ePub

2026-05-14 · source · source ↗ · raw capture · wiki page

By Robin Whittleton (December 11, 2025)

Core Argument

ePub uses HTML technology but with significant departures from web development. Developers must "unlearn" conventional knowledge to work effectively with ePub's constraints.

Key Technical Differences

XHTML Foundation

ePubs require XHTML based on the XML specification, not modern HTML5: - Syntactically valid XML markup required - Self-closing tags mandatory - Correct namespace declarations - XML attributes (xml:lang)

"XHTML didn't work out" for web browsers due to fragility and performance issues, yet persists as the ePub standard.

CSS Limitations

E-readers use "positively historic" rendering engines. Developers should avoid modern pseudo-classes like :is() and :not(). CSS must account for XML namespaces using the pipe separator.

"I'm wary of using :not() in ePub CSS for a widely distributed title."

MathML and SVG

Both require namespace declarations in XHTML with prefixed elements.

ePub-Specific Semantics

The epub:type attribute enables semantic richness unavailable in standard HTML (endnotes, backlinks). Gradually being deprecated in favour of Digital Publishing WAI-ARIA spec.

Extended Vocabularies

Z39.98-2012 Structural Semantics Vocabulary extends ePub through namespaced attribute values.

ePub File Structure

META-INF/container.xml pointing to package files
Package file with metadata, manifest, and spine
XHTML files for content
Zipped archive with .epub extension

Recommends Standard Ebooks toolset for beginners.