FAQ

A family of systems that together form a semantic substrate for software. Each system owns one idea and builds on the ones beneath it; the goal is that software written later can safely use software written earlier, through evidence the substrate can check. Start with the vision and the architecture.

The substance stack (ontos → arche) and the evidence loop are built and running; the dashboard tracks each member's real channel, release, wiring, conformance, and docs. The examples mark what's real today versus still a sketch, and what ships today lists runnable commands. Capability and realization are candidate theory, not settled.

Because the committed ledger is a bare, reproducible snapshot that doesn't read the member repos — the live site regenerates the observed view at deploy time. See Interpret a dashboard absence.

Most of the site is a projection of atlas's generated data — topology, the docs catalog, ADRs, the evidence ledger — imported directly, never copied. The authored prose on top is the one part that can drift, so it's linted (atlas site lint) for real system references and resolving links. The generated figures are gated by design figure lint.

Docs are each member's own exported documentation. Theory is the exploratory substrate-theory track — pre-normative reasoning, not specification. Decisions are the settled cross-cutting ADRs. Concepts and terms tie them together.

See the runbooks — run the checks locally, add site content, add an example, trace a dependency, write a cross-repo ADR. Everything is zero-dependency on the atlas side; only the site has npm dependencies.

Yes — atlas and the site are built openly; the constellation lives at github.com/Bitspark.