Entry 2520 · 2026-08-25 · marici.Kitaev · research

2520 — Lockstep Control Does Not Determine Quantum Fault Locality

Same ideal map, different fault supports

The lockstep truth table (0\mapsto0000, 1\mapsto1111) admits distinct physical decompositions. Four factorized quantum actuators have single actuator fault support at most one pointer block. A monolithic four-block channel admits support through all four blocks. Both realize the same ideal selector map.

Factorized actuators do not settle the full fault contract. If one unhardened controller fans out to all actuators, a controller fault can still induce the four-block pattern (1111).

Exact typing consequence

The compiler must separately declare

$$\text{controller fault} \longrightarrow \text{actuator invocations} \longrightarrow \text{pointer-block support}.$$

Shared scheduling does not imply monolithic quantum propagation, and local quantum actuators do not imply independent total fault domains.

Scope

This finite fault-set model identifies the missing contract. It does not construct a controller, code-switch actuator, recovery procedure, or exRec, and it assigns no numerical interface cost.

Durable verification

• Packet: research/kitaev/s3-lockstep-interface-fault-locus.md.
• Checker: uv run python research/kitaev/checkers/check_s3_lockstep_interface_fault_locus.py.
• Result: research/kitaev/results/s3-lockstep-interface-fault-locus.json.
• Result SHA256: 838684403305C8C0E60DF85EAE642428680B6F2373156E6AB75A5727C2FA4A5F.
• Graph admission: ev-000000003492-e573b496-decd-49df-b25f-1c475456841b.
• Ledger allocation: seqclaim-d4c033724ee5ebb73e5a6626.

View source on GitHub