Skip to content

Style and glossary

The words your documentation uses, enforced by the linter.

Two files hold the editorial decisions your team already makes in review. glossary.yaml names the words. style.yaml names the voice.

terms:
- use: project
not: [space, site]
because: A project is one documentation property. The app says "project".
- use: workspace
not: [org, organisation, tenant]
because: A workspace holds the people, the billing and the projects.
usestringrequired

The word this documentation uses.

notstring[]required

The words it does not use. Each one becomes a lint warning wherever it appears.

becausestring

Why. It appears in the lint message, so the writer who hit the rule learns the reason instead of only the correction.

A word from a not list on a page raises wrong-term, a warning. It is a warning rather than an error because a quotation may legitimately contain the other word.

voice:
- Use the active voice.
- Write short sentences. One instruction to a sentence.
- Say what the product does today, not what it will do.
avoid:
- Do not use "simply", "just" or "easy".
- Do not promise a feature that is not built.

Two lists of short rules. At most 15 of each, and each at most 160 characters.

The limit is deliberate. A rule a person can hold in their head is a rule a model can follow, and a paragraph of guidance is neither.

The lists travel with every request for help with the words. A suggestion comes back sounding like the rest of your pages rather than like a model.

They are the sentences you already say in review. Writing them down once means you stop saying them.

RuleSeverityWhat it means
glossary-invaliderrorThe YAML does not match the schema.
style-invalidwarningThe YAML does not match the schema.
wrong-termwarningA page uses a word the glossary rejects.

The glossary is an error and the style guide is a warning because the glossary is read by the linter on every page, and a broken one would silently stop enforcing your vocabulary.