The Integrity Engine.
Deterministic, cross-language data integrity detection in 0.068ms.
Traditional hashing (SHA-256) tells you that a payload changed, but it cannot tell you how it changed. In financial ledgers, a benign key reordering triggers the same alarm as a catastrophic float-truncation bug.
LatticeUVW2x maps canonical JSON payloads to the S² sphere and computes the geodesic angle (drift) between them. It localizes the mutation, classifies it with a confidence score, and does it in sub-millisecond time.
The Benchmark
Independent testing against standard cryptographic hashing.
| Mutation | SHA-256 Detects? | Holonomy Detects? | Drift (rad) | Holonomy (ms) |
|---|---|---|---|---|
| float_truncation (ERP rounding) | ✅ | ✅ | 0.864688 | 0.0684 |
| unicode_drift (NFC ↔ NFD mismatch) | ✅ | ✅ | 2.000373 | 0.0231 |
| key_reordering (benign) | ❌ | ❌ (ignored) | 0.000000 | 0.0169 |
Developer Interface (PyO3 / Rust)
Enterprise partners can query the integrity kernel directly via the Python bridge. The Rust core does the rest.
from routing_holonomy import py_validate_holonomy
original = '{"amount": 100.00, "currency": "TRY"}'
analytical = '{"amount": 100, "currency": "TRY"}' # float truncation
verdict = py_validate_holonomy(original, analytical)
print(verdict["drift_angle"]) # 0.864688
print(verdict["is_identity"]) # False
Mathematical Guarantees
We do not hide behind marketing terms. The kernel makes 5 mathematically provable invariants (Determinism, Intactness, Change Detection, Range Invariant, and HALT).
We also explicitly document what we do not guarantee (such as cross-language Go/Rust hash equality, which is tracked for Phase 6). We believe honest scoping is the highest form of trust.