DONE / RELEASED: three native routes for whole-flow cancellation after #4222

[brianross93]

DONE / RELEASED: three native routes for whole-flow cancellation after #4222

Status / UTC update / priority revision: DONE / RELEASED / 2026-10-20 19:30 UTC / revision 7.

Outcome (LOCAL-ONLY): all three bounded tests are complete. The retained-side consumer preserves every reachable packet degree of freedom at the pinned h=100; a fixed common native representation repays the saving at the full output boundaries; and an exact port-specific common factor is blocked by the compressed gather's lost information, with the tested full-port expansion already losing its recursive cost comparison. Small exact native/character and algebraic checks passed. These are scoped results, not a general impossibility or an improved selected multiplication bound. Variable/mixed native encodings remain open. The reports/checkers remain local and are not publicly reproducible from this post. No new PR, artifact upload or paid compute occurred; no independent human review is implied. This scope is released; #4222 remains released.

Question: can the actual three-stage complex motif's reachable gather relation remove paid recursive work through (1) consumer-first output contraction with chronological sources, (2) a shared physical representation, or (3) a common native core plus lower-cost structured corrections? These are three bounded tests of one whole-flow proposal, with one boundary and cost ledger, not a catalogue or a broad search.

Predecessor #4222 remains DONE / RELEASED. Its clean-increment relation is a starting hypothesis, not a native saving. All three tests retain the complete signed bank-exchange target, arbitrary dirty auxiliary outputs, exact address/operator order and paid acquisition, storage, routing, decoding, guard bits and disposal. Compare against clean-private-centre execution, already present in the selected lineage; do not recount the older three-to-two centre rewrite as a new gain. Any small upstream instance is a mechanism test, not automatic evidence for the selected caller.

Fresh pin: canonical main 12fa8c5c844be21bfefb38a19503fbd867a11244; selected conditional kappa remains 0.001672660197076918. Sources: upstream/build/sections/03-motifs.tex, 02-streams.tex, 05-layers.tex; retained selected clean-centre and native owner-route contracts. The precise applicability of the scalar relation will be checked before use. Required outcome per route: a changed paid-child identity and preliminary full cost, a concrete obstruction in the tested family, or a clearly inconclusive result.

Overlap refreshed: centre/shared/joint/cross-stage discussion and open-PR searches plus recent #1166 handoffs. Main and selected record are unchanged. The active PR3147 independent sink/braid/circle audit, GL3 normalization audit, Gaussian finite-core work, QQ16 and D4 are excluded. No known active same-scope claim appeared in this bounded coverage; please flag overlap.

Completed within the original 60-minute budget using small exact calculations. No work remains reserved by this claim.

Brian Ross supplied the conceptual strategy: normalize the complete known computation, propagate input relations forward and output requirements backward, and predict cancellation before producing expensive terms. He directs this three-route follow-on; analysis and implementation use OpenAI Codex assistance. Prior construction and obstruction authors retain their credit.

{"schema":"integer-mult-coordination/v1","kind":"DONE","task":"three-native-routes-after-4222","scope":"Consumer-first, shared-representation and common-core tests of the same upstream three-stage complex flow, against clean-private-centre execution, retaining full dirty/native boundary and paid costs. Excludes P112 audit, GL3, QQ16, D4 and Gaussian work.","base":"12fa8c5c844be21bfefb38a19503fbd867a11244","expiry":null,"artifact":null}