Skip to content
Sasha Lopashev edited this page Jul 19, 2026 · 44 revisions

BHCP

Beyond Human-Centric Programming (BHCP) is an experiment in making intent, authority, limits, and required evidence first-class program meaning. A person states what must be true and what may happen; machines remain free to discover an acceptable execution without silently weakening the contract.

BHCP is early but executable. The repository is not claiming a complete language, planner, runtime, or SDK.

Current maturity

  • Implemented: a safe-Rust canonical-source slice; typed canonical AST and semantic IR; deterministic CBOR; SHA3-512 algorithm-tagged identities; CDDL validation; a minimal reducer kernel; self-hosted all, homogeneous-output any, none, typed causal chain, and unary typed gate with complete algebra conformance; registered-verifier/evidence-bundle foundations; a capability-bounded local verifier process runner; deterministic process-result mapping into targeted evidence bundles; a fail-closed reproducible coding-agent experiment controller; monotonic layered policy with scoped, expiring, auditable waivers; and the audited bounded presentation-profile layer.
  • Measured: six checked-in coding-agent pilots plus the pre-registered multi-seed comparison. The strongest frozen case shows canonical policy order preventing one plausible prose-only error; Pilot 006 adds an honest negative skill arm and a successful follow-up. The valid five-session repeat produced five unchanged 4/10 candidates and 0/5 acceptance with correctly conservative claims; the broader evidence remains mixed and explicitly limited.
  • In progress: Phase 2 now advances through in-session registered adapters #27.
  • Planned: complete the Phase 2 evidence-loop audit #28 after #27.

Start here

The repository remains the source of truth: README, vision, normative semantics, CDDL bundle, and conformance catalog.

Clone this wiki locally