[amaybaum]

The fixed-rule ring can transport a supplied setting request into a common carrier preparation, but the inspected OI sources do not supply a physical producer of its complete ready/request precursor. This paper confirms that source gap after checking the existing apparatus work, broader repository sources and primary reversible-computation literature. It makes no new impossibility claim.

Current head: 450fbdfefcf1b37de831c63c3943a64a1ec57cc9, from main b1d782f954f02cec578a10e071b73e48f86eb3dd. Only verification/programmes/oi-qm/reconstruction/physical-controller-20261107/LEAD.md is added (blob ca4a61d585025a959eab909de3f6abbbe011eda1), through two normal single-parent commits. Accountable home #866; owner scope. Author controller_lead; separate reviewer controller_review.

Independent exact-head ACCEPT AS A BOUNDED SOURCE-GAP RECORD, with no blocking findings. The physical controller gate is NOT PASSED; no further candidate derivation follows. The localized quiet-neighbor requirement is specific to the screened loader, not every possible controller. Common source law, controller knowledge, full output/failures and absence of automatic reset remain explicit.

Ordinary verification 42414185362 completed successfully at this head: 25 successful jobs and one conditional skip, verified from structured run/job records. This is repository CI, not a V3 F/E attestation or evidence of physical availability.

SA-1 still supplies P3/E-iso. Native A, the directional nonlinear ring and Q-A's reciprocal linear update are kept separate. Q-A #1184 at cfe5200d4aa78647ba7d290d59664c590cdf5a14 is unreviewed compatibility provenance only; no theorem or author self-check from it is consumed. Any transfer requires the actual update, physical loading/timing/readiness and full observer history without parity exposure during reuse/recombination. #865 drive and #1147 readiness remain open; their work is not activated here.

Paper-only source selection is complete. No simulation, enumeration, implementation, new physical-premise adoption, freeze, special dispatch, merge or manuscript promotion. Draft retained for the reviewed record; no next construction is authorized by this closeout.

[amaybaum]

Independent source-selection review at 450fbdfefcf1b37de831c63c3943a64a1ec57cc9, base b1d782f954f02cec578a10e071b73e48f86eb3dd.

Verdict: ACCEPT AS A BOUNDED SOURCE-GAP RECORD. No blocking findings. The physical controller source gate is NOT PASSED; this is not ACCEPT-FOR-PAPER-DERIVATION. No controller availability, new physical premise, Q-A execution, or permission to proceed beyond the selection stop is established.

I reviewed the entire one-file delta: physical-controller-20261107/LEAD.md, blob ca4a61d585025a959eab909de3f6abbbe011eda1, 378 added lines. The comparison contains exactly this addition and two single-parent commits from the stated base. The PR head still matched the reviewed SHA immediately before this review. AGENTS matches blob 496eb7018c736a064dddcf00fd1a278790872f49.

1. The stop is justified at its stated scope. The selected ring is a concrete preexisting state-dependent physical example, rather than a newly supplied target gate. Its binary specialization has one fixed F(u,v)=(f(u) XOR v,u) for both settings, with f_i=u_(i-1) XOR (u_(i+1) OR u_(i+2)). However, the maintained loader assumes the complete precursor root and eight hidden constraints. Its forward identity transports a supplied request and correlated readiness; it does not construct their producer. Thus the first missing physical premise is an accounted source of (5)–(6), including selectable Y_1=c with the same uncontrolled eta. Quiet X_2=X_3=0 is a useful localized sub-obligation for this selected loading route, not a universal necessary condition for every future controller. The packet correctly calls this a confirming source-selection stop, not a new obstruction or an exhaustive physics search.

2. Common preparation, controller knowledge, failures and reuse are retained. The hidden precursor law is shared across requests; changing only Y_1 changes only the loaded program coordinate after the loading tick. This is conditional independence in the supplied realization, not physical selectable loading. The complete declared controller has C_1=(Z,c), C_2=(c,Z), so receiver masking at c=1 does not hide Z from the observer. Possible earlier preparer knowledge is correctly implementation-dependent. The 1/4 load-match probability and 2^-8/2^-12 event masses are limited to the stated canonical source and are not factory or herald rates. The failure complement, correlations and full outgoing state remain; Z=1 and c=1 give the stated failures of the old readiness contract. No reset, persistent log, physical clock, fresh bath, or compatible next instrument follows.

3. The general OI sources do not close that gap. I inspected Main §1.2 and the rooted-family audit: the fixed prior/root/projection architecture is positive realization structure, with counting a selected baseline and structured priors allowed. Main's action-labelled theorem serves all action sequences in one constructed realization, but its explicit action writes, ledgers and clock do not implement an actuator in another fixed rule. The ReadWriteControl definition/theorem and InstrumentDilation statement match the packet: selectable permutation families and the supplied seed/control matrix identity are their actual interfaces. Computation/book explicitly condition their BQP model on preparation, control, clock, measurement, precision, noise and operational closure. The packet neither denies these results nor promotes their hypotheses to native hardware.

4. Primary literature supports the comparisons, not an OI controller. I read the cited scans and relevant sections. Bennett 1973, pp.525–527, constructs the reversible three-tape computation with blank auxiliary storage and retained input/output. Fredkin–Toffoli 1982, §§6.1–6.4, supplies an actual idealized collision/routing mechanism with prepared inputs, mirrors, geometry and synchronization. Margolus 1984, §3 and §§7–9/Appendix A, distinguishes arbitrary-function second-order reversibility from universality of particular rules and signal layouts, including the specified mod-3 neighborhood construction. No embedding of those mechanisms into the selected ring, wave or Native A is provided. Their universal-computation capability therefore does not satisfy this source gate. No thermal cost is inferred.

5. Model and verification boundaries are correct. SA-1 explicitly supplies P3/E-iso, the controller schedule and clock; its common meter supply is preserved. A6's transformation of link data is not their physical evolution. SM's directional nonlinear OR ring does not supply Q-A's reciprocal symmetric linear arming. Native A's different rule/resources/access are not spliced into either model. The current continuous-substates standing explicitly limits its paper proofs and author-reported reviews; the packet uses those contracts as context rather than independently verified premises. Existing positive, conditional and negative studies remain intact, and no global no-go is inferred.

6. Q-A compatibility is an obligation, not a result. The off-main cfe5200d draft is used only to identify its proposed update/access contract; I do not accept or re-review its mathematical claims or self-checks here. Any later transfer needs an explicit common realization executing the required symmetric/NNB and translation updates, actual setting loading, supplied readiness/known-one resources, and the complete controller/preparer record cut. Parity inaccessibility must survive that actual access through reuse/recombination, and the next operation must accept the full correlated output. Declaring the parity hidden is insufficient. Posit I is not adopted.

This was independent source/type and paper-scope inspection. I ran no scientific computation, enumeration, simulation, Lean build, implementation or dispatch and made no author-file edits. Ordinary PR CI is separate repository verification and does not change this scientific disposition. Further candidate derivation remains blocked pending a concrete physically sourced controller and a new independent gate decision.

[amaybaum]

Owner normal-merge approval: #866 c.6718471099. Previous #1196 post-merge main CI 42457511542 is green. Current main c06eb8098a5ff587ffb9c35151c8cd3dd5015595; head remains reviewed 450fbdfefcf1b37de831c63c3943a64a1ec57cc9, independent source-gap ACCEPT and ordinary CI 42414185362 green. Raw host PR data confirms mergeable. API confirms merge base b1d782f9; a conflict-free merge using that base gives tree 61dcd86ddf2b2c893005b5bfb3a1ce7d70ddc9cd. Its delta versus current main is byte-identical to the original reviewed one-file/378-line delta. Source/proof boundaries and non-supporting off-main Q-A context retain their reviewed scope. Proceeding with ordinary merge; wait for green post-merge main CI before #1177. #866 remains open; the physical source gate remains NOT PASSED.

[amaybaum]

Normal merge 6a987a0ba9898f2e0a9115dd5854a04725473078 has the verified predicted tree and both expected parents. Post-merge main CI 42458124176 completed successfully: 25 successful jobs and one conditional skip. The integrated report remains a bounded source-gap record; controller preparation/source and Q-A transfer are not established. #866 remains open. Proceeding only to the next authorized integration; CP-PREP-1 remains gated until all five integrations are green.