Skip to content

Commit 9544fe6

Browse files
hyperpolymathclaude
andcommitted
docs(proofs): scope remaining obligations TG-3/5/6/7/8
Add per-obligation scoping (approach, effort, risk, dependencies) to the registry so the roadmap survives context compaction. Recommended order: TG-3 (keystone; reduces to OCaml-infer ≡ Lean-infer via TG-2) → TG-5 (cheap property test) → TG-7 (braid word problem) → TG-8 (mechanical) → TG-6 (research-grade WASM bisimulation). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
1 parent 6f55e71 commit 9544fe6

1 file changed

Lines changed: 53 additions & 0 deletions

File tree

PROOF-NEEDS.md

Lines changed: 53 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -67,6 +67,59 @@ Cross-referenced to [PROOF-NARRATIVE.md §3](PROOF-NARRATIVE.md#3-remaining-obli
6767
For full per-obligation statements, _why valuable_, and the
6868
assumptions each rests on, see PROOF-NARRATIVE.md.
6969

70+
## Scoping of the remaining obligations (2026-06-14)
71+
72+
Concrete approach, effort, risk, and dependencies for what is left after
73+
TG-0/1/2/4/9/10 landed. **Recommended order: TG-3 → TG-5 → TG-7 → TG-8 → TG-6.**
74+
75+
### TG-3 — OCaml `typecheck.ml` refines Lean `HasType` *(keystone, recommended next)*
76+
- **Key lever:** TG-2 already proves Lean `infer ≡ HasType`. So refinement
77+
reduces to **OCaml `infer_expr` agrees with Lean `infer` on the shared core
78+
fragment** — a cross-language differential check, not a fresh metatheorem.
79+
- **Approach:** translation-validation harness. Generate core-fragment terms,
80+
type each with OCaml `infer_expr` and with Lean `infer`, assert equal types
81+
(and equal accept/reject). Seed from existing corpora + a small generator.
82+
- **Honest boundary:** the OCaml checker is far larger than the 26-rule core
83+
(match, weave, compute, cap/cup, mirror/reverse/simplify, twist, pipeline,
84+
two-pass program typing, width inference, Bool-eq). Deliverable = refinement
85+
validated on the core fragment + an explicit "extra-core" feature list, each
86+
marked model-later or declare-non-core. A universal Lean proof is out of reach
87+
without modelling the OCaml program itself.
88+
- **Started:** `Eq` aligned to `tEqWord` this session. **Effort:** ~3–5d.
89+
**Risk:** med (scoping the fragment). **Deps:** TG-2.
90+
91+
### TG-5 — `compositional.ml` rewriter preserves types *(cheap, do early)*
92+
- **Approach:** OCaml property test first — random compositional exprs →
93+
`compile` → assert PD invariants (arc balance, crossing-count = unit-word
94+
length, closedness, residue = verbatim source for `EchoClose`). Lean model of
95+
the IR + a preservation theorem is a later, optional second rung.
96+
- `test_compositional.ml` now exists (the old "no test file" blocker is gone).
97+
- **Effort:** property test ~1–2d; Lean model ~3d. **Risk:** low. **Deps:** none.
98+
99+
### TG-7 — `eqBraids` decides braid-group equivalence *(high-value domain work)*
100+
- Current `Step.eqBraids` is list equality = a **soundness floor**, not
101+
completeness (misses σᵢσⱼ=σⱼσᵢ for |i−j|≥2 and σᵢσᵢ₊₁σᵢ=σᵢ₊₁σᵢσᵢ₊₁).
102+
- **Approach:** implement Dehornoy **handle reduction** (practical braid word
103+
problem) in OCaml with tests against known equivalences/trefoil; mechanize
104+
correctness (Garside/Dehornoy theory) in Lean later.
105+
- **Effort:** OCaml ~1w; Lean proof ~2w+ (deep). **Risk:** med-high (the maths).
106+
**Deps:** none (pure domain algorithm).
107+
108+
### TG-8 — each dialect is a conservative extension of core *(mechanical, voluminous)*
109+
- 5 dialects (braid-calculus, quantum-circuit, skein-algebra, string-diagram,
110+
virtual-knot). Conservativity = core `HasType`/`Step` unchanged; new rules
111+
fire only on new syntax.
112+
- **Approach:** model + prove ONE dialect end-to-end as a template, replicate.
113+
- **Effort:** ~1w each (+1w for the first/template). **Risk:** low. **Deps:** the
114+
dialect specs/impls must exist.
115+
116+
### TG-6 — WASM compilation preserves semantics *(research-grade, plan separately)*
117+
- **Approach:** practical first rung = **differential testing** (source `eval`
118+
vs wasm exec over a corpus, assert equal observable outputs). Full
119+
bisimulation proof (WasmCert / Wasm-spec) is a multi-week research project.
120+
- **Effort:** differential ~1w; proof ~3w+ research. **Risk:** high (proof),
121+
low (differential). **Deps:** a runnable wasm runtime to exec against.
122+
70123
## Proof categories
71124

72125
| Code | Meaning | Applies? |

0 commit comments

Comments
 (0)