[DLamdata]

This extends #2564’s scalar F0→F2 quotient to a conditional native source-response factorization for the selected prefix: 11 complete local sweeps, four bridges and four idle sweeps, with centre ablations at calls 0 and 10.

For the complete prefix F, ablated diagnostic A, initial X0 source data d, and arbitrary other contents h held fixed:

E_D d = (F-A)(d,h) - (F-A)(0,h)
P_D d = first eight complete native pre-transport centre-field responses to (d,0)

ker(P_D) ⊆ ker(E_D), hence E_D = L_native P_D.

Signed boundary identities cancel the X0 response at B1. At F2, all eight pre-centre rows vanish on the 112 target and 99 helper roles that remain potentially live. The proof compares complete terminal continuations.

Start with the commit-pinned proof and scope and validation record. Only research/centre-native-terminal-factor/ changes.

The result assumes the declared complete native common-frame and cover/local-view interface. The eight fields include owner, payload and spectator indices. It does not establish sharp native rank eight, recover the dirty offset, supply a paid decoder, rebuild the physical compiler, reduce the invoice or change κ. A fully charged dirty-residual consumer remains open.

Local validation

• All three registered checks passed in 100.148 seconds on CPython 3.12.14, including exact signed boundary identities and an independently derived reverse-adjoint check.
• The three mathematical checkers also passed with -I -O -B in a relocated, read-only export without .git. The Git-dependent metadata launcher and committed-byte checker ran separately in the Git checkout. Corruption controls detected altered sources and receipts.
• Committed-byte verification passed for 22 package/input files; all 12 published package blobs matched the reviewed commit. Detailed commands and resolved review findings are in the validation record.

Remote CI — snapshot at 2026-10-16 10:29 UTC

Fork checks: 56 successful / 23 failed. Upstream PR checks: 61 queued.

All failed fork jobs were inspected. Observed causes were inherited source/receipt-pin drift, singleton-target unpacking, missing historical Git objects, ledger/digest mismatches and symlink-cycle handling. Exact guards, targeted parent/head reproductions and checkout-history evidence established their presence in integration parent ba3de472…. No contribution-related failure was found. Blocked later steps and queued upstream checks remain unverified; no complete CI pass is claimed.

Comparison with upstream 475fec190d5ad746bd4690574dca30d6901acdaf found scientific pins, relevant guidance and package verification tools unchanged. No rebase or mathematical rerun was warranted. Reviewed head: f122d255c4cdb0714c079d8e91b76bd7a34da23e.

Credit ywz65/#2564, the cited lineage (#2048, #2199, #2391, #2461), and Sagar Thakrar’s unchanged Clifford kernel. Prepared by DLamdata with substantial Codex assistance. The checks share the selected archive/native-view premise; primary and boundary checks share the kernel. AI review is not independent human or formal global verification. New work is Apache-2.0.

Canonical Discussion update.