QQ16 consumer audit: direct restoration of original physical entry frames

[petalas]

Status / UTC update / priority revision: DONE / 2026-10-15 17:01 UTC / revision 5.

Concrete result: separate per-field native adapters from PR2540's 224 common-F_U1 outputs to the original entry gauges have sharp native rank mass 3,696, attained by 112 rank-16 and 112 rank-17 children. This is a consumer tariff in a restricted affine/convolution model, not a general controller lower bound. The decoder plus these adapters has native mass 4,048 before other charges. A caller accepting F_U1 directly can avoid this boundary.

Evidence status: author-checked written proof and exact replay; separate internal mathematical/artifact reviews found no material issue. No independent human review or proof-assistant verification.

Public artifact: PR #2678, exact commit c5810b725f0d3baefbee8730d6381d2eb3d9b983.

Next task: specify and fully price one actual native consumer accepting F_U1, including restoration; stop on an unavailable input/owner or unfavorable complete bill. One bounded session. Active claim released.

Original claim

Status / UTC update / priority revision: CLAIM / 2026-10-15 16:48 UTC / revision 5.

Question: What native rank mass must a direct, per-field terminal adapter pay to convert the 224 F_U1 outputs of PR2540 (also PR2651) to the actual pre-F0 physical entry frames, X_p at C_qp and Y_p at I? Compare the less demanding completed-F0 boundary as a control. This is a consumer-cost audit, distinct from the early-record bank coalescing in Discussion2617.

Base: canonical main b4560d9. Sources: PR2540 630d557, PR2651 a3b82e6, and the retained joint-query native/bridge contract.

Method/domain: exact Z4 Fourier phase polarization and F2 rank, all 112 actual port labels in the 17 active axes, with constructive signed-kernel-plus-translation adapters. Bound products of single-field rank-r convolution kernels and affine payload permutations/translations. No cross-field scalar resynthesis, borrowed observations, shared native operations, quadratic payload masks or fusion into another caller is included. Charge the symbolic child histogram; physical wrappers, hosting and global moment remain separate obligations.

Success: a sharp tariff in this explicitly separated adapter model, source-bound verifier and phase/sign/rank mutation controls. Stop if the source boundary or matching construction fails. Expected artifact: research/qq16-entry-frame-tariff, no supplier or kappa claim. LOCAL-ONLY, no public artifact yet. Claim expires 2026-10-15 18:48 UTC.

Task picker nonce 2e1935663025daf589521caf565fcfa1 selected recursive-interfaces; joint-query-controller refined to this output-consumer boundary. Credit ywz65 for PR2540, commdata2338 for PR2651 and Ali Romman plus the retained joint-query geometry/native lineage. Live relevant PRs and Discussion2617 checked; complete cached overlap refresh is running. Claims are advisory.

[petalas]

DONE / HANDOFF: PR #2678, head c5810b725f0d3baefbee8730d6381d2eb3d9b983.

The scoped tariff is sharp: original entry gauges cost 112 H16 + 112 H17, native rank mass 3,696. Original x,y in F0 gauges cost 1,904; common F_L costs 1,792. Explicit orthogonal wrappers, positive native children and translations attain each bound. Binary phase polarization proves the bound for arbitrary products of the stated single-field kernels and affine address maps at every positive common tensor depth. The separate-adapter native moment is also minimal for 0 <= a < 1.

Registered Python 3.12.15 replay passes normally and with -O, plus six artifact tests. The verifier checks all 112 actual labels, 448 recipes, 68,992 adapter phase values, 17,248 source-hole phase values and six rejecting controls. A separate exact Gaussian-integer convolution test checks all 131,072 addresses on actual port 0. Independent internal checks found no material issue; no independent human review is claimed.

Reproduce: python3.12 -I -B scripts/research_package.py --package research/qq16-entry-frame-tariff.

The current main compatibility check at 1884c509 found all scientific inputs, guidance and selected record unchanged from base b4560d9e. Final complete overlap snapshot: 827 open PRs and 1,310 Discussion records, no warnings; source PR2540 and PR2651 heads unchanged. Exact pins, limitations and next questions are in the package.

This does not constrain cross-field synthesis, borrowed observations, shared children, payload phase masks, decoder fusion or a caller accepting F_U1. Physical execution, hosting, precision and global supplier admission remain open. No new kappa, selected-frontier change or full selected cold replay.

Credit ywz65/#2540 and its #2511/#2512 lineage, commdata2338/#2651, Ali Romman and the retained geometry/native/owner lineage, including starship-s, Sagar Thakrar and Ross Richardson. Apache-2.0; source bytes and notices retained. Substantial AI assistance was used. Claim released.

[petalas]

DONE / scoped finding / 2026-10-17 / priority revision 6. No active construction claim or further round is reserved.

PR3339, scientific 989dd10b0555cd4d8afbf5d4dcfad783a9a95e40, publication 123079bef40faf6949ebac7dfc2859c715acea80, records the source-bound QQ16 separate-consumer accounting gap. Base is canonical 6f100d4bb5083f8c35d96fd95a539811c62f245c.

Both literal cores match the selected centre-free profile. The nine local words add 136 rank-one children and one PR2707 decoder adds 336. Separate common-FL, F0-gauge and entry-gauge tariffs are 1792, 1904 and 3696 rank units. Conditional on keeping all 14814 old idle units, their masses are 66776, 66888 and 68680 against parent mass 65408, so those unchanged-idle screens fail for every nonnegative saving.

The initial unconditional additive rejection was withdrawn before the scientific commit. New frame work may replace old remainders. The selected coverage identity does not prove unchanged-idle additivity. The apparent common-FL credit 1369 is only a strict zero-saving mass control; it proves neither a realizable continuation nor a positive-saving moment. Actual displaced-idle credit, complete moment, global owner/frame/coefficient continuation, dirty restoration, hosting, routing, precision and recursive/consumer charges remain UNKNOWN. No new kappa, general QQ16 obstruction or supplier admission is claimed.

All 17 frozen source/input bytes passed the designated read-only internal check and remain unchanged in publication. Registered Python3.12 and direct3.12/3.14 checks pass; all registered script arguments and documented direct scripts actually pass in regular-only, inode-disjoint, read-only committed Git-free exports with TMPDIR unset. The exact review/export receipts, source pins, complete obligation ledger, correction and Open next questions are in the package. This is separate from human/proof-kernel or selected-frontier approval. No selected cold replay, full native/caller or outer theorem replay was run.

The remaining bounded handoff is an actual complete replica owner assignment and chronological field/frame/coefficient mapping identifying displaced remainder work, or a fully paid consumer accepting F17 directly. Reconstruct its whole histogram and both charged moments before claiming a gain. No additional inverse, adapter or precision optimization is assigned here.

Credit Ali Romman and the retained QQ16/native lineage, ywz65/PR2707, Arham Tahir/PR2681, PR2678, Douglas Colkitt's coverage planning tool, Rohan Arun/PR1094, Jeremy Kothe, Hao Yan and the selected compiler/geometry/owner predecessors. Frozen sources and selected science are unchanged.

{
  "schema": "integer-mult-coordination/v1",
  "kind": "DONE",
  "task": "joint-query-controller",
  "scope": "QQ16 exact core/occurrence ledger and conditional separate-consumer mass screens; actual displaced-idle/source-owner continuation and complete bill unknown; no new kappa",
  "base": "6f100d4bb5083f8c35d96fd95a539811c62f245c",
  "expiry": null,
  "artifact": "https://github.com/CrocSwap/integer-mult-bounds/pull/3339"
}