PDF and EPUB glossary

Short definitions of the terms you meet when you convert a PDF to EPUB and read it on an e-reader. Each one says what the term is in two sentences and why it matters for the conversion.

Updated: Nov 4, 2026

Inside an EPUB

A reflowable EPUB: A reflowable EPUB stores text and its structure, not positions on a page, so the reading app lays it out again for each screen, font and size. It is the default kind of EPUB and the opposite of a fixed-layout one.
A fixed-layout EPUB: A fixed-layout EPUB keeps each page exactly as designed, like a PDF, and is declared with rendition:layout set to "pre-paginated" in the package document. It suits picture books, comics and cookbooks where the position of every element matters.
EPUB 3: EPUB 3 is the current version of the EPUB e-book standard, built on XHTML5, CSS, SVG and an HTML navigation document. Its latest revision, EPUB 3.3, became a W3C Recommendation in 2023.
EPUB 2: EPUB 2 is the previous generation of the EPUB standard (2.0.1 dates from 2010), based on XHTML 1.1 and an NCX table of contents. It has been superseded by EPUB 3 but many older files still use it.
The EPUB container (OCF): An EPUB file is a ZIP archive laid out by the Open Container Format (OCF): an uncompressed "mimetype" file first, a META-INF/container.xml that points to the package document, and the content files. Renaming a .epub to .zip lets you look inside.
The OPF package document: The OPF (Open Packaging Format) file, or package document, is the XML file that describes an EPUB: its metadata, the manifest of every file and the spine that sets the reading order. Reading apps find it through META-INF/container.xml.
The EPUB manifest: The manifest is the list, inside the OPF package document, of every file that belongs to the publication, each with an id, a path and a media type. A file in the ZIP that is not in the manifest, or a manifest entry with no file, is an error.
The EPUB spine: The spine is the ordered list, inside the OPF package document, of the content documents a reader pages through from start to end. It sets the default reading order of the book.
The NCX table of contents: The NCX (Navigation Control file for XML) is the table of contents format of EPUB 2, an XML file with the media type application/x-dtbncx+xml. EPUB 3 replaced it with the HTML navigation document but still allows an NCX for older readers.
The EPUB navigation document: The navigation document is the XHTML file that holds an EPUB 3's table of contents, as a nav element with epub:type="toc", and optionally landmarks and a page list. Every EPUB 3 must have exactly one, marked with the "nav" property in the manifest.
EPUB landmarks: Landmarks are an optional nav element in the EPUB 3 navigation document (epub:type="landmarks") that points to key places such as the cover, the table of contents and the start of the main text. Some reading apps use them for "Go to beginning" and similar commands.
An EPUB page list: A page list is a nav element (epub:type="page-list") in the EPUB 3 navigation document that maps the page numbers of a printed edition to places in the e-book. It lets readers cite or jump to "page 42" of the print source.
XHTML in EPUB: EPUB content documents are XHTML: HTML written with XML syntax, so every tag must be closed, attributes quoted and the file well-formed. EPUB 3 uses the XHTML syntax of HTML5; EPUB 2 used XHTML 1.1.
CSS in EPUB: EPUB uses ordinary CSS stylesheets for presentation, but reading apps may override fonts, margins and colours, and some properties, such as fixed positioning, are unreliable or not allowed. Good EPUB CSS is short and lets the reader decide.
Epub:type: epub:type is an attribute that adds meaning to an element in an EPUB 3 document, using terms from the EPUB Structural Semantics Vocabulary, such as "footnote", "noteref", "chapter" or "pagebreak". Reading apps use it, for example, to show a footnote as a pop-up.
A media type in EPUB: A media type (MIME type) such as application/xhtml+xml, image/jpeg or text/css tells the reading app what each file in an EPUB is. Every manifest item declares one, and it must match the file's real content.
Dublin Core metadata (dc:*): EPUB metadata uses Dublin Core elements in the package document: dc:identifier, dc:title and dc:language are required, and dc:creator, dc:date, dc:publisher and others are optional. EPUB 3 also requires a dcterms:modified timestamp.
An EPUB cover image: The cover image of an EPUB 3 is the manifest item marked with the "cover-image" property; reading apps show it as the thumbnail in the library. It is usually a JPEG or PNG of the front cover.
Reading order: Reading order is the sequence in which text is meant to be read: column by column, with side notes, captions and footnotes placed where they belong. In an EPUB it is set by the spine and the order of elements in each document.
EPUB core media types: Core media types are the file formats every EPUB 3 reading system is expected to support, so they need no fallback: GIF, JPEG, PNG, SVG and WebP images, MP3, AAC and Opus audio, CSS, TrueType, OpenType, WOFF and WOFF2 fonts, and XHTML. EPUB 3.3 lists them in a table.
A foreign resource (and its fallback): A foreign resource is a file whose media type is not one of EPUB's core media types, such as a TIFF image. EPUB 3.3 lets you include it only with a fallback, declared in the manifest with the fallback attribute or provided by HTML, that a reading system can show instead.
The EPUB unique identifier: The unique identifier is the primary identifier of an EPUB: the value of the dc:identifier element that the package element's unique-identifier attribute points to. It is often an ISBN or a UUID, and EPUB 3.3 says a significant revision needs a new one.
Dcterms:modified: dcterms:modified is the required last-modified date of an EPUB 3, written in the package document as a meta element. EPUB 3.3 requires exactly one, in the form YYYY-MM-DDThh:mm:ssZ (UTC).
Remote resources in EPUB: A remote resource is a file an EPUB loads from the web instead of carrying it inside the ZIP. EPUB 3.3 allows it only for audio, video, fonts and resources fetched by scripts; everything else, including images and stylesheets, must be inside the container.
Scripted content in EPUB: A scripted content document is an EPUB content document that contains JavaScript or an HTML form. EPUB 3.3 allows scripting, but the manifest item must carry the scripted property and reading systems are free to run the scripts in a restricted way or not at all.
Non-linear content in EPUB: Non-linear content is a document listed in the EPUB spine with linear="no": auxiliary material, such as notes, answer keys or descriptions, that is outside the main reading order. Reading systems may skip it while paging, so it has to be reachable from a link.
Font obfuscation: Font obfuscation is a simple scrambling of embedded font files, keyed to the EPUB's unique identifier and declared in META-INF/encryption.xml, so a font cannot be copied out of the ZIP and installed elsewhere. EPUB 3.3 says it is not encryption and offers no real protection.
A synthetic spread: A synthetic spread is two adjacent pages shown side by side on one screen, like an open book. In a fixed-layout EPUB the rendition:spread property says when a reading system should do it: none, landscape, or both orientations.

Inside a PDF

A PDF text layer: The text layer of a PDF is the actual character data (which glyph, in which font, at which position) behind what you see on the page. You can select and search text only when it exists.
A scanned PDF: A scanned PDF holds photographs of paper pages, so each page is an image with no text layer unless OCR was run on it. It looks like text but a computer sees only pixels.
OCR (optical character recognition): OCR (optical character recognition) turns images of text into characters a computer can read. Applied to a scanned PDF it adds a text layer, with recognition errors that depend on scan quality, fonts and language.
A tagged PDF: A tagged PDF contains a structure tree that marks headings, paragraphs, lists, tables and the reading order, on top of the drawn text. Tags are what screen readers use, and PDF/UA requires them.
A ToUnicode map: A ToUnicode map (CMap) is a table in a PDF font that says which Unicode character each glyph represents. Without it, a program can draw the text but may not know what the letters are.
An embedded font: An embedded font is a font file (often only the glyphs used) stored inside a PDF or an EPUB, so the text displays the same on any device. In EPUB it is declared with @font-face in the CSS.

Text and typography

A running header: A running header (or footer) is text repeated at the top or bottom of every page of a printed document, such as the book or chapter title and the page number. It is page furniture, not part of the text.
A footnote in an EPUB: A footnote is a note printed at the bottom of the page it refers to, linked to the text by a reference mark. In an EPUB there is no page bottom, so the note becomes a linked element that readers can show as a pop-up.
An endnote: An endnote is a note collected at the end of a chapter or of the book instead of at the foot of the page. In EPUB 3 it is marked with epub:type="endnote" inside a section of endnotes.
A noteref: A noteref is the link from a reference mark in the text (a superscript number or symbol) to its note, marked in EPUB 3 with epub:type="noteref". Reading apps that support it open the note in a pop-up.
A ligature: A ligature is a single glyph that joins two or three letters, such as "fi", "fl" or "ffi". Fonts use them for appearance; the underlying text should still be the separate letters.
A soft hyphen: A soft hyphen (U+00AD) is an invisible character that marks where a word may be split; it only shows as a hyphen when the line actually breaks there. In a PDF, a hyphen at the end of a line can be a soft hyphen or a real one.

E-readers and delivery

Sideloading: Sideloading means putting an e-book on a reader yourself, by USB cable, email or an app, instead of buying it in the device's store. Converted EPUBs always reach a reader this way.
Send to Kindle: Send to Kindle is Amazon's service for delivering your own documents to a Kindle library, through an app, a web page or a personal Kindle email address. It accepts EPUB files and converts them to Amazon's format for the device.
KF8 (AZW3): KF8 (Kindle Format 8) is Amazon's e-book format based on HTML5 and CSS, introduced in 2011; files in this format usually have the .azw3 extension. It replaced the older MOBI format on Kindle devices.
KEPUB: KEPUB is Kobo's variant of EPUB, recognised by the .kepub.epub file extension, which Kobo devices open with their own rendering engine instead of the generic EPUB one. Kobo's own store uses it for the books it sells.
DRM (digital rights management): DRM (digital rights management) is encryption that ties an e-book or PDF to an account or device so it cannot be opened elsewhere. Store-bought e-books are often protected this way.
EPUBCheck: EPUBCheck is the official, open-source validator for EPUB files, maintained by the W3C and released under the BSD-3-Clause licence. It reports every problem as a message with a code such as RSC-005 and a severity: fatal, error, warning, info or usage.
An EPUB reading system: An EPUB reading system is the software that opens an EPUB and presents it to a reader: an e-reader's built-in app, Apple Books, a Kobo or a phone app. EPUB 3.3 defines it as a system that processes EPUB publications for presentation to a user.
An EPUB CFI: An EPUB CFI (Canonical Fragment Identifier) is a standard way to point at any location or range inside an EPUB, written as epubcfi(…) with a path through the package and the document. Reading systems can use it for bookmarks, highlights and reading positions.
MOBI: MOBI is the older Mobipocket-based e-book format that early Kindles used, with the .mobi extension. Amazon's Send to Kindle help page no longer lists it among the supported formats, while it does list EPUB.

Accessibility, languages and media

EPUB accessibility metadata: Accessibility metadata are schema.org properties in the package document, such as schema:accessMode, schema:accessibilityFeature and schema:accessibilitySummary, that describe how accessible an EPUB is. Stores and libraries show them to readers who rely on assistive technology.
Alt text: Alt text is a short text alternative for an image, written in the img element's alt attribute, that screen readers read aloud in place of the picture. Decorative images get an empty alt.
Page-progression-direction (RTL books): page-progression-direction is an attribute of the EPUB spine (ltr, rtl or default) that tells the reader in which direction pages turn. Arabic, Hebrew and Persian books use rtl, and so do Japanese books set vertically.
Vertical writing in EPUB: Vertical writing sets text in columns from top to bottom, read right to left, as in traditional Japanese, Chinese and Korean books. EPUB 3 supports it through the CSS writing-mode property (vertical-rl).
MathML: MathML is the XML language for mathematical notation, allowed inside EPUB 3 content documents so equations stay text instead of images. Support in reading apps varies.
SVG in EPUB: SVG (Scalable Vector Graphics) is the XML vector image format allowed in EPUB 3, either as image files or inline in content documents. Unlike JPEG or PNG, it stays sharp at any size.
EPUB media overlays: Media overlays synchronise recorded narration with the text of an EPUB 3, using SMIL files that pair audio clips with elements of the content documents. Reading apps highlight each phrase as it is read aloud.
A language tag (dc:language and xml:lang): A language tag is a code such as en, en-US, ar or ja that says which language a text is in, written with the BCP 47 syntax. An EPUB declares it in dc:language in the package document and with lang and xml:lang on its content documents.
A page-break marker: A page-break marker is an element in an EPUB, usually with epub:type="pagebreak" and the page number as its label, that records where a page of the printed edition began. Together with a page list it lets readers find "page 42" of the print source.
WCAG: WCAG (Web Content Accessibility Guidelines) is the W3C standard for accessible web content, with success criteria at three levels, A, AA and AAA. Its current version is WCAG 2.2, and EPUB Accessibility 1.1 builds on it.
EPUB Accessibility 1.1: EPUB Accessibility 1.1 is the W3C Recommendation, published in October 2024, that defines when an EPUB is accessible and how to say so. It requires discovery metadata, WCAG 2 conformance and EPUB-specific features such as page navigation where a print edition exists.
Ace by DAISY: Ace by DAISY is a free accessibility checker for EPUB from the DAISY Consortium, available as a command-line tool and a desktop app. It helps evaluate an EPUB against accessibility practice; it does not replace a human review.