Document header

The starred group at the top of every file — the title, description and tags a document carries.

#Writing the header

A document opens with a starred attribute group — * at the start of a line, followed by ( ). It is the first thing in the file, it carries no body, and it is the one construct written with the * sigil.

*(
    title: The 68–95–99.7 rule, live
    description: However you set k, the band from mu - k sigma to mu + k sigma holds the same share.
    tags: statistics, probability, normal-distribution
)

scene{
    …
}

Inside, it is an ordinary attribute group: key: value pairs, one per line or separated by semicolons, exactly as any other group is written.

#What each field does

The header is not a fixed schema. It is a bag of attributes the compiler carries through to whatever consumes the document, and each consumer reads the keys it cares about. For an internote those are:

title — the document's name, shown wherever it is listed or linked.
description — one sentence, used for previews and search results.
tags — a comma-separated list, used for subject browsing.

A key the consumer does not recognise is carried without complaint rather than rejected, which is what lets one compiler serve documents with different metadata needs — the newsroom articles on this site read date, category and excerpt from the very same construct.

class is reserved. The compiler consumes it to decide which element library the document is compiled against, so it never reaches the header attributes a consumer sees.

#Where the header may appear

At the top of the document, once. A partial spliced in with @include is source, not a document of its own, so it carries no header — the metadata belongs to the file that is being compiled.

Everything else about a document's shape is scene and below; Grammar covers how the rest of it is written.