coeval

coeval is a base for declarative languages. Each language keeps its programs as JSON documents. A document gets its meaning from a question, not from a definition: its kind names the evaluator, and a ladder finds it.

The site explains each idea with a live diagram. The diagrams use the model package of this repository, thus each answer comes from real code.

Status

The kernel is on npm as @mark1russell7/coeval. Its API is not stable before version 1.0. Since version 0.2, each language evaluates through the kernel (decision D13). The plan is in docs/ROADMAP.md, the decisions are in docs/DECISIONS.md, and each rule has an ID in docs/SPEC.md. docs/SEMANTICS.md gives the same rules as a small calculus of inference rules.

The model in seven rules

1. A program is a document. It is a value in an envelope: { "type": "add", "value": [2, 3] }. The type names a kind.
2. A question finds the meaning. The answer comes from the own entry of the document, then from its kind, then from the default.
3. An evaluator evaluates its parts, and does not look inside them. Thus equal parts can replace each other.
4. A context is the only door. A term reaches outside itself through five roots: self, names, world, event and caps.
5. Order has a meaning only for effects. Pure work runs in any order that obeys its data. Effects keep their written order.
6. A reference is a path from a root of the context. Each step is a question to the thing that the path reached. A type is a path into the world.
7. A language evaluates through the kernel. The kernel compiles a document one time into a plan, checks it before a run, and evaluates the shared constructs. A language gives its operations and its own rules.

Packages

Package	Contents
@mark1russell7/coeval	The kernel: envelopes, the ladder, kind contracts, sealed parts, plans, the run side with capabilities, error values with a place, the static kit, the formula kit and a small standard world
@coeval/site	The site: a visual grammar where each concept has one glyph, a guide that adds glyphs one page at a time, and a playground. GitHub Pages serves it.

Commands

pnpm install
pnpm typecheck                       # the source and the tests of each package
pnpm test                            # the tests of each package
pnpm lint:ste                        # the writing rules
pnpm --filter @coeval/site dev       # the site, with the dev server of Astro
pnpm package add <name> --preset=ts  # make a new package

Writing style

The prose of this repository follows ASD-STE100 Simplified Technical English (STE). STE is a style target. ASD does not certify this repository. The project glossary is in ste.config.json. Keep a local copy of the STE dictionary in .ste/, and do not commit it.

Disclosure

An AI model (Claude, from Anthropic) wrote most of the text and the code of this repository, under the direction of the author. The STEMG of ASD-STE100 asks for this disclosure in its white paper on STE and artificial intelligence (June 2026).

License

MIT. Refer to LICENSE.