EPUB CFI & ANNOTATION SYNC

Highlights that disappear when a reader switches devices. Bookmarks that reopen at the wrong paragraph. Last-read positions that jump backward every time a book is opened.

These are not rare edge cases. They are what happens when annotation systems are built without a solid understanding of how EPUB actually anchors content.

EPUB CFI & Annotation Sync is a practical production handbook for engineers building reading applications, annotation platforms, and ebook distribution systems. It explains the EPUB Canonical Fragment Identifier from first principles: what a CFI is, how it is generated and resolved, why it fails, and how to design systems that remain reliable under real-world conditions such as reflow, DOM changes, content updates, offline use, and multi-device synchronization.

Inside, you’ll find:

The complete CFI grammar: steps, assertions, offsets, indirection, and ranges Generation and resolution algorithms with working code Stability analysis: what actually invalidates a CFI — and what does not Production-ready data models for highlights, bookmarks, and last-read positions Cross-device synchronization using G-Set CRDTs and LWW Registers Offline-first architectures with outbox queues, retries, and backoff Conflict-resolution policies designed around real user expectations Storage, indexing, and query strategies capable of scaling to 100 million annotations Testing techniques including round-trip, property-based, and chaos testing A complete case study of a production-grade annotation service

This is not an introductory tutorial. It is written for engineers who already ship software and need to understand the trade-offs, failure modes, and architectural decisions behind robust EPUB annotation systems.

Every chapter ends with a production checklist. Technical claims are grounded in the EPUB 3.3 specification, and areas where the specification leaves behavior undefined are clearly identified.

If your reading system loses highlights, misplaces bookmarks, or forgets where the reader stopped, this book explains why — and shows you how to build it correctly.