Conformance and correctness harness
Started: public stable-core corpus and runner. Next: CI and diagnostic snapshots.
Optimization plan
Make the language reliable before making it loud.
The full roadmap is also tracked in GitHub under docs/OPTIMIZATION_PLAN.md.
Native backend reliability
Gate x86_64 Windows/Linux with deterministic smoke tests before widening claims.
Standard library hardening
Keep stable core small, documented, and tested.
Developer experience
Diagnostics, formatter, lint, LSP, and examples must feel predictable.
Ecosystem, Python bridge, C ABI, benchmarks, releases
Production replacement claims wait for repeatable proof.