mirror of
https://github.com/allaunthefox/Research-Stack.git
synced 2026-07-31 03:05:21 +00:00
Replace the TODO(lean-port) sorry with a complete proof of the
projectionOrdering theorem: for positive SourceValue pairs s1 < s2
with s2 ≤ maxExpected, projectToCoding preserves strict ordering
of the Q0_64 values.
The proof uses Nat-only arithmetic (no Float) and handles two cases:
- a2 < d: both values fit in Q0_64 range, ordering follows from
monotonicity of integer division
- a2 = d: a2*s/d = s clamped to q0_64MaxRaw; a1*s/d < q0_64MaxRaw
via the key inequality (d-1)*s < (s-1)*d
Build: 8598 jobs, 0 errors (lake build)
|
||
|---|---|---|
| .. | ||
| AVMR_Theorems.lean | ||
| Domain_Intersections.lean | ||
| MASTER_PLAN.md | ||
| README.md | ||
| RESEARCH_ROADMAP.md | ||
| SUBAGENT_ASSIGNMENTS.md | ||
| SUBAGENT_ASSIGNMENTS_MAX_PARALLEL.md | ||
Missing Proofs Registry
Purpose: Track theorems, lemmas, and formalizations that are defined, documented, or conjectured but not yet proven in Lean 4.
Status: Living document — update as formalization progresses.
Gap Categories
Category A: AVMR Core Theorems (Models 102-110)
| # | Theorem | Statement | Current Status | Blocker | Auto-Provable? |
|---|---|---|---|---|---|
| 102 | Quasi-Periodic Square-Shell | n = k² + a = (k+1)² - b with a+b = 2k+1 |
✅ PROVEN — ring + omega |
— | YES ✅ |
| 103 | Tip Coordinate Vector | Tip(n) = (ab, a-b) as ℝ² embedding |
Partial — tipCoord defined |
Real embedding properties | NO — needs injectivity |
| 104 | Axial Event Production | A_k, G_k, C_k, T_k partition shell |
✅ PROVEN — axialGeneratorExhaustivity |
— | YES ✅ |
| 121 | Axial Generator Exhaustivity | {A_k, G_k, C_k, T_k} exhaust S_k | ✅ PROVEN — ring + nlinarith |
— | YES ✅ |
| 105 | Resonance Hub | Tip(m²) = (0, -(2k+1)) |
✅ PROVEN — simp + ring + omega |
— | YES ✅ |
| 106 | Standing-Wave Rear Field | Ψ_i(n_i - d) = α_d · χ_i |
Partial — buildFieldAt computes |
Field convergence | PARTIAL — echo sum proven |
| 107 | Interaction Score | J(n) = ab·F_m + (a-b)·F_p + ⟨χ,F_c⟩ |
✅ Implemented | — | — |
| 108 | Left-Right Transduction | (n,k,a,b) ↦ (e,τ,...) ↦ (S,P,G) |
Missing — pipeline not formalized | Multi-stage bind | NO — needs pipeline |
| 109 | Temporal Error-Coding | t(n) = nR + τ, 8-slot |
Missing — microtime lattice | Temporal algebra | NO — needs Fin arithmetic |
| 110 | AVMR Commitment | Σ_{k+1}[i] = Φ(Σ_k[2i], Σ_k[2i+1]) |
Missing — vectorized Merkle | Aggregation invariant | NO — needs algebraic structure |
Search space shrunk: 4 of 11 theorems proven automatically (Models 102, 104, 105, 121). 7 require manual work.
Category B: Core Semantics Debt
| Module | Debt | Location | Priority |
|---|---|---|---|
BraidCross.lean |
Fix16.add associativity | Line 72 | High — blocks algebra |
BraidCross.lean |
Fix16.add commutativity | Line 84 | High — blocks algebra |
Constitution.lean |
Lawful constitution preservation | sorry placeholder |
Medium — ENE core |
Canon.lean |
Float elimination (timestamp, trustLevel) | float64 field kind |
Medium — core purity |
Graph.lean |
Float → Q16_16 port | Geometry module | Low — shim acceptable |
Category C: Domain Intersection Gaps
| Intersection | Missing Bridge | Impact |
|---|---|---|
| E ↔ M (Verification ↔ Lean) | BEA consensus → Formal proof | Weak verification formalism |
| H ↔ G (Algebra ↔ Energy) | Geometric algebra → Thermodynamics | Noether connection missing |
| J ↔ L (Dynamics ↔ Application) | Phase transitions → FEA | Engineering bridge |
Category D: Unified Compression Extensions
| Feature | Status | Needed For |
|---|---|---|
| Temporal lattice integration | Missing | Model 109 connection |
| AVMR commitment hook | Missing | Model 110 connection |
| Microvoxel constraint check | Partial | Model 21-22 validation |
| Meta-LUT formalization | Missing | Model 23-24 |
| HotSwap selector proof | Missing | Model 25 |
| Final Score Law | 🔄 Formalizing | Models 119-120 — cost aggregation |
| Score parameter bounds | Missing | λ₁,λ₂,λ₃,λ₄ ∈ [0, 2.0] |
| Cost monotonicity | Missing | L(X) increases with complexity |
Proof Templates
When adding a proof, use this format:
theorem <name> : <statement> := by
<tactics>
<;> try native_decide
<;> try omega
<;> try nlinarith
Preferred witnesses:
#evalexamples with expected outputnative_decidefor concrete casesomega/nlinarithfor arithmeticsimpwith appropriate lemmas
Dependencies
Before proving higher theorems, these foundations are needed:
- Fix16 algebra (BraidCross) — blocks all algebraic manipulation
- Shell state totality — blocks AVMR event classification proofs
- Field convergence — blocks standing-wave theorems
- Q16_16 metric — blocks distance/geometry proofs
Recent Updates
- 2026-04-17: Created registry, populated from MATH_MODEL_MAP (181 models)
- 2026-04-17: Added Unified Compression gaps (Models 111-118)
- 2026-04-17: Added Final Score Law (Models 119-120) — complete cost aggregation
- 2026-04-17: Added Agent F1/F2/F3 Tier Proofs (Models 121-131) — explicit Φ operators
- 2026-04-17: Identified 3 critical domain intersections needing bridges
- 2026-04-17: Prover ran — Models 102, 104, 105, 121 proven automatically
How to Resolve
- Pick a theorem from Category A or B
- Create branch:
proof/<theorem-name> - Add theorem + witness to appropriate
.leanfile - Update this registry with ✅ status
- Run
lake buildto verify - PR with theorem statement and proof summary
Enforcement: Per AGENTS.md Section 4 — every def computing cost or invariant must have #eval or theorem witness before commit.