Reading and revision

Three tools that read the manuscript rather than set it. None of them changes the book unless you tell it to, and none of them appears in an exported file.

Readability

Scores and highlights for the chapter you are in.

The gauge

A Flesch reading-ease score, with three bands marked:

Band	Score	Typical of
Essay	around 20	Academic and literary non-fiction
Literary	around 45	Literary fiction
Mainstream	around 70	Commercial fiction, most non-fiction

The bands are reference points, not targets. A dense literary novel sitting at 30 is doing what it means to.

What is measured

Words, reading time, words per sentence, syllables per word. Under Pacing: the share of characters inside quotation marks (dialogue), words per paragraph, and sentence count.

Highlights

Switch on Highlight in the manuscript and the categories you have enabled are marked in the text itself:

Category	Flagged when
Very hard sentences	Five separate readability formulas agree it is above about grade 14
Hard sentences	The same five agree it is above about grade 10
Complex words	Bureaucratic register with a plain equivalent (utilize, ascertain), offered alongside
Adverbs	Usually a sign a stronger verb exists
Passive voice	Names no one doing the action
Filler	Intensifiers and hedges the sentence is stronger without (very, rather, kind of)
Repetition	The same word three times within fifty words

A sentence is flagged only if it is at least 14 words long and all five formulas agree. A short sentence is never called hard, however dense — the formulas are unreliable at that length, and a panel that flags “He ran.” teaches you to ignore it.

Consistency

Finds where the book does the same thing two ways.

The panel opens on what it is measured against — Itself, or one of the guides written for the book’s language — and, against itself, by saying what it is not: nothing is called wrong here; this is only what the book does two ways. Burned and burnt are both correct — a book that uses each of them fifty times is the problem.

Under the headline, one card per rule with something to settle: its name, how many places would change, and what they become. Typical findings: spelling variants (analyze / analyse), the serial comma present and absent, numbers under a hundred in words or in figures, repeated spaces, inconsistent dashes, a quotation opened and never closed.

Open a card and it shows the two habits with how often the book does each, then walks that rule’s occurrences one at a time: the sentence, the change written into it, and Fix this one under it. A quotation is shown whole — opening mark, the words, closing mark — rather than one end at a time, and the occurrence is marked in the manuscript while the card is open. A line at the foot says where the work is: Mostly in chapter seven, 40 of 42 is a chapter pasted from elsewhere, and you can press it to go there.

A card whose rule the book does both ways about equally says so — Choose — and offers the two habits. Pick one, and the other places are offered as changes.

Fix all 42… on the card takes the whole rule, and Fix all 162… beside the headline takes every rule at once. Both open the list before anything changes: every sentence that would, chapter by chapter, with the change written into it, and Apply at the bottom. Untick the ones you want left alone — or skip a whole chapter — and press the arrow beside a change to go and look at it in the manuscript. It is one edit, and ⌘Z takes all of it back; the history calls it Fixed 162 style points.

Ignore sets a card aside for this sitting. Already consistent at the foot lists the rules the book keeps to one habit already, with that habit named.

Looking before deciding

The eye beside each habit walks every occurrence of it in the book, the ones that already conform included — every serial comma, every burnt. Nothing is offered on a stop that needs no change. Press the eye again to stop.

Your decisions travel with the book

The habit you settle on, and what the book is measured against, are saved in the book itself. They are the same on your other devices and ⌘Z takes them back like any edit. Which cards you ignore is not saved: that is a way of looking, not a decision about the manuscript.

Style guide

Pick a guide instead of Itself to check against a named manual rather than the book’s own habits. Each card then says which guide asks for the change.

Language	Guide
English	Chicago Manual of Style (18th edition), New Hart’s Rules (2nd edition)
French	Lexique des règles typographiques en usage à l’Imprimerie nationale (2002)

Use Itself to make the book agree with itself, and a guide when a publisher has asked for one.

Quotation marks

Converts typewriter marks to typeset ones across the whole book. It sits with the other typographic decisions rather than among them: it is not a habit the book has two of, but one change the whole book takes at once.

Typed	Becomes
'	’ — apostrophe or single quote
"	“ “ — opening and closing double quotes
-- or ---	— an em dash; between two numbers, -- becomes an en dash (1914–1918)
...	… — an ellipsis

The dialog lists what it found and what it will change, grouped by kind, before you convert anything.

An apostrophe and a closing single quote are the same character, which is why a straight ' cannot be converted correctly by rule alone in every case. The dialog shows the first line in the book that would change, before and after, so you can check what it decided — and the conversion is one edit, so ⌘Z takes the whole of it back.

Version history

Every save on this device, grouped by day — today, yesterday, then dates. It opens from the save indicator because it answers the same question one step further on: the indicator says the work is safe, and this is where you see it.

Most rows are just a record of what happened: the edit is already in the book, so there is nothing to bring back and no button.

A row does carry Re-apply when it holds a version that was set aside rather than committed. That happens when the same block was edited on two devices: the remote version wins, and yours is preserved here instead of being lost. The count on the menu item — History · 2 kept — is how many are waiting, and a dot on the save indicator says there are some before you open it.