@some-ui/hangul-game-core
v0.0.9
Published
Pure-Rust game engine for content-typing games, compiled to WASM. Hosts the challenge/matcher/difficulty machinery consumed by `@some-ui/honeycomb`.
Readme
hangul-game-core
Pure-Rust game engine for content-typing games, compiled to WASM. Hosts the
challenge/matcher/difficulty machinery consumed by @some-ui/honeycomb.
Architecture decisions live in docs/adr/; the derivations
those ADRs state as given live in the canons below.
[!IMPORTANT] > Canon-governed workspace — read the canon before editing this package.
docs/canon/hangul-progression-canon.typ— The Single-Glyph Ceiling — the content-model algebra this crate implements: stimulus, answer sequence, token cursor, content domaindocs/canon/adaptive-learning-canon.typ— The Unobservable Learner — why the engine'sHashSetof completed identities and its single shared difficulty scalar cannot become adaptive (Prop. 2.2, Thm. 6.1), and why the!show_romanizationgate must stay (Cor. 3.2)These are not background reading. They are the documents this package is derived from: modules here are checked against a Definition / Axiom / Theorem number, not against a feature spec. If a change cannot be traced to a canon citation, either it belongs somewhere else or the canon is missing an amendment that should land first.
Human reviewers: a diff that changes behaviour governed by a canon and cites nothing is incomplete — ask for the citation. LLM agents: read the cited sections before proposing a change, and never silently renumber or rewrite a canon result. See
docs/canon/README.mdfor the amendment discipline.
