Adding a glossary and an index to a book

Write a glossary: it is a sorted list of terms, it takes an afternoon, and it works in every format. Only build a page-referenced index if the book will be printed and readers will consult it out of order — in a reflowable ebook page numbers do not exist, so the index becomes a list of links or nothing at all.

These two get mentioned in the same breath and they are not the same kind of job. One is a list you write. The other is a professional craft with its own trade body, and the reason your book probably should not have one is that the format most people publish in cannot support it.

The difference, and which you need

GlossaryIndex
What it isTerms and their definitionsConcepts and where they appear
Sorted byAlphabetAlphabet, with sub-entries
EffortAn afternoonDays, or a professional's fee
Works in an ebookYes, unchangedOnly as links
Needs final paginationNoYes, for print
Worth it forAny book with jargonReference books people consult

The glossary, which you should probably write

Cheap, genuinely useful, and it does something for the book beyond the reader: writing one forces you to notice the twelve terms you used without ever defining, which is the most common failure in a first non-fiction draft.

  1. Collect the terms as you do the line-edit pass — anything a reader outside your field would not know. Twenty to fifty is normal; two hundred means you are writing a dictionary.
  2. One or two sentences each. A glossary entry is not an encyclopedia entry. If it needs a paragraph, it needed a section in the book.
  3. Define it as you use it, not as the field defines it. If your book uses "edition" in a narrow sense, say so.
  4. Sort alphabetically. Thematic grouping defeats the point — nobody browses a glossary.
  5. Cross-reference sparingly with "see also", and only where the other term is genuinely needed to understand this one.

In Markdown it is a definition-style list, which survives both exports cleanly:

## Glossary

**Bleed**
Artwork extended past the trim line so that a cut
slightly off-target leaves no white edge.

**Front matter**
Everything before the first chapter: title page,
copyright, contents, dedication.

**Reflowable**
Text that adapts to the reader's screen and type
size, rather than sitting on a fixed page.

Bold term, definition underneath, blank line between entries. It reads correctly as plain text, sets correctly in the PDF, and needs no special handling in the EPUB. Put it in the back matter, before the about-the-author page — front and back matter has the running order.

Why a real index is hard

Because an index is not a list of where words appear. It is a list of where *ideas* are discussed, which requires reading the book and making judgements a search function cannot make.

  • It indexes concepts, not strings. A discussion of pricing that never uses the word "pricing" still belongs under pricing.
  • It excludes passing mentions. A term named once in an aside is noise in an index, and automated tools cannot tell the difference.
  • It needs sub-entries. "Pricing, psychology of" and "Pricing, and refunds" are the structure that makes an index usable.
  • It depends on final pagination. Every page number is only correct for one specific layout, so the index is the last thing made and breaks if anything before it changes.
  • It is a profession. Indexers train for it and charge accordingly — commonly a few dollars per printed page, which on a 300-page book is a real number.

And why it does not survive an ebook

A reflowable EPUB has no pages. The reader picks the type size, and the text flows to fit — so "see page 74" is meaningless, and a converted index becomes a column of numbers pointing at nothing.

FormatWhat an index can be
PrintPage numbers. The real thing
Fixed-layout PDFPage numbers, and they work
Reflowable EPUBLinks to sections, or nothing
KindleLinks, and the device's own search

Which gives the honest answer for a digital-first book: the search function is your index. It is better than an index at finding a string, worse at finding a concept, and it is what readers will actually use. PDF or EPUB covers why the two formats differ on this and much else.

The middle option: a linked index

If the book is genuinely a reference and you want something in the ebook, build a short index of links to sections rather than pages. Twenty to forty entries, each pointing at a heading. It is not a professional index and it is considerably better than nothing.

## Index

- **Bleed** — see [Print covers](#print-covers)
- **Cover sizes** — see [Covers](#covers), [Print covers](#print-covers)
- **EPUB validation** — see [Exporting](#exporting)
- **Margins** — see [Page setup](#page-setup)

This works in both exports: the links are tappable in the EPUB and clickable in the PDF, and neither depends on pagination. It is also maintainable, which a page-referenced index is not — the anchors survive a change of trim size, a change of font, and a second edition.

If you do want a print index

  1. Finish the interior completely first. Final text, final trim size, final type settings, final margins. Nothing after this point can change.
  2. Export the final PDF and work from it, because the page numbers in that file are the ones going in the index — print-ready PDF.
  3. Hire an indexer if the book matters commercially. Ask for a sample of two chapters before committing to the whole thing.
  4. Or do it yourself from the contents outward: go heading by heading, list the concepts discussed under each, then sort and merge. Slow, but tractable at 150 pages.
  5. Keep it in the manuscript, not in a separate file, so the next edition starts from something rather than nothing.
  6. Re-check it after any change. One added paragraph in chapter two moves every page number after it.

That last point is the whole economics of the thing, and the reason the linked version is the right answer for most self-published books. An index is a snapshot of one specific layout; everything else in a plain-text manuscript is cheap to change precisely because it is not — see correction, revision or second edition? for what that costs you later.

The short recommendation, then. Write the glossary — it takes an afternoon, it improves the book, and it works everywhere. Add a linked index if the book is a reference. Commission a real index only if you are printing a book people will consult on a desk, and only once nothing else can move.

Read next

Guides · 5 min read

Front matter and back matter: what goes in a book

Front matter is everything before chapter one: half title, title page, copyright page, dedication, epigraph, contents, and sometimes a foreword or preface — in that order. Back matter is everything after the last chapter: appendices, notes, glossary, acknowledgements, an about-the-author page, and whatever you want the reader to do next. Neither is decoration; both are conventions readers use without noticing.

Read it
How-to · 5 min read

How to add a table of contents to a Markdown book

A printed table of contents is a page you write: a list of chapter titles, in order, with a page break before the first chapter. An EPUB does not need that page — its table of contents is generated from the heading structure and lives in the reader's navigation. Get the headings right and both fall out of the same manuscript.

Read it
Explainers · 4 min read

EPUB or PDF: how to choose, and why you probably want both

Use PDF when the page is fixed and matters — print, handouts, workbooks, anything with a layout the reader must see as you set it. Use EPUB when the text should adapt to the reader's screen and type size, which is every e-reader and almost every ebook store.

Read it

Write it in Markdown. Ship a book.

bukpress turns the file you are already writing into a typeset PDF and EPUB — your type, your covers, no watermark on any plan.

Start free — no card required