The quiet interface: how this site is built

The design rule that governs every surface on this site — every brand, every dashboard, down to a single button — is written in the project's own configuration, and it is enforced before anything is called finished:

Nothing exists unless it is load-bearing. Before adding a border, an icon, a label, a shadow: if it were removed, would meaning, function or clarity be lost? If not, it does not go in.

That single rule explains almost every visible decision below.

When new material comes in, it is not typed into a WYSIWYG editor. It is shaped into a Story Document: a typed list of blocks, each with a stable id so the Studio can address it directly and each field can carry up to nine languages.

The schema is deliberately wide. A case study might use `steps`, `figure` and `compare`. A retrospective might reach for `timeline`, `chart` and `bento`. A press note might just be `lead`, `stats`, `callout`. The point is not that every story uses every block — it is that the site never needs a new component to say something new; it composes from the same 26.

The guide on this site, the search index, the schema.org markup and even the plain-text file made for other AI systems all draw from the same source: a single truth file, where every fact carries a grade — externally verifiable, or self-described and plausible. Anything that could not be verified is written down too, but kept in a pending list that is never rendered.

This is slower than writing marketing copy. It also means the guide can be asked anything, in any language, and it will either answer with a citation or say plainly that it does not know.