SOURCE_AUDIT.md

Source and novelty audit — Matter Embryogenesis v3

All source research was read from the user's available files. Inputs are identified by exact filename and SHA-256 in provenance/sources.json. Relevant documents and the EDD proof supplement are retained under provenance/source_documents/. No scientific claim is accepted solely from a filename, search snippet, or a previous project's maturity score.

User research actually integrated

Source	Inspected content	Transferred object	Explicit limit
TRUE MACHINE MEMORY RCCE v1	Local recoverability model, quadratic proof, complementarity, liability ledger, limitations	Evidence/control value and the cost of destroying recovery resources	A loss-plus-effort identity, not a probability of successful chemical repair
EDD Joint Reserve v2	Manuscript and supplement: conditional-mean decision proof, matched subspaces, Hölder onset, scalar/matrix resource frontier	Matched sensor/repair resource bundles and a paid reserve threshold	Precision is not bytes; mobility is not channel radius; fixed conditional actuator assumption matters
Evolvability Closure v5	Continuation grammar, finite-rank closure/minimality, identification, limitations	Distinguishing present response from future manufacturing state	No low predictive rank or efficient planning inferred for matter
RAVEN v4 certificate protocol	Complete protocol and trust-boundary warning	Witness dependencies and explicit verifier assumptions	A metadata passed flag is not a formal proof or reliable physical signal
3-D conductivity audited v6	Scope, response state, continuity, cubic construction, exact-volume rounding, calibrated physical distance	Actual physical response versus abstract admissibility; surrogate-to-binary error accounting	Infinite hierarchy is not an efficient finite compiler; homogenized response is not automatically a finite module port response

These are user-supplied research releases, not independent peer-reviewed validations. The needed finite proofs were independently reconstructed here. The full 3-D hierarchy and cognitive claims are not premises of R1–R9.

Primary literature checked in this revision

• Dorfler and Bullo, arXiv:1102.2950: abstract and full PDF, including the Schur response formula and boundary-current interpretation.
• Doyle and Snell, arXiv:math/0001057: author-posted text, especially Thomson/Dirichlet principles around PDF pages 49–50.
• Stern et al., PRX 11, 021045 / arXiv:2011.03861: primary abstract, establishing local physical coupled-learning prior art.
• Dillavou et al., Physical Review Applied 18, 014040 / arXiv:2108.00275: primary abstract/publisher record, establishing an electronic variable-resistor laboratory demonstration.
• Altman et al., Physical Review Applied 22, 024053: primary publisher abstract for elastic-network coupled learning.
• Spielman and Srivastava, arXiv:0803.0929: primary record for spectral-approximation prior art.
• Korman, Kutten, Peleg, DOI 10.1145/1073814.1073817: primary conference record for local certification; follow-up primary proof-labeling records also inspected.
• Teng et al., Nature Communications 16, 3238, DOI 10.1038/s41467-025-58422-0: primary publisher record, supported by the prior full source audit; later publisher fetches intermittently failed. First author corrected to F. Teng. The result uses external patterning.
• Brookhaven CFN primary account: silica/tin-oxide framework device integration. This is laboratory/publisher evidence of selected material routes, not a demonstration of the proposed architecture. Source.

The v1 audit retains primary checks for algorithmic assembly complexity, intrinsic universality, proofreading tile sets, experimental algorithmic DNA assembly, inorganic framework templating, and transport/thermodynamic prior art. They were not represented as new literature discoveries in v2.

Claim classification

Established ingredients: Schur/Kron reduction; Dirichlet and Thomson energy principles; Loewner spectral approximation; concentration/union bounds; conditional-mean quadratic control; Hölder inequality; local acknowledgements and proof-labeling ideas; self-assembly programming and proofreading; selected DNA-to-inorganic routes.

Reconstructed restricted derivations: R1 response nonaccumulation; R2 finite measurement/final-seal envelope; R3 finite bounded-increment reachability; R4 finite-horizon confidence; R5 postorder access; R8 energy envelope; R9 weighted fault concentration.

Candidate contribution in this package: R6's combined response-certified maturation statement and its developmental compiler interpretation; use of recoverability liabilities for material closure; a paired local-function/transport/closure implementation that replaces v1's unsupported device fault-fraction contract.

Imported prior user result applied to fabrication: R7 quadratic closure liability and the EDD reserve threshold. These are not relabeled as discoveries first made in this report.

Engineering proposals: a compatible temporary service/controller layer, bounded reserve deposition after conversion, local electrical isolation, molecular acknowledgements, and physical multi-material compilation.

Not established: independent priority; a universal chemical alphabet; molecular implementation of all primitives; macroscopic performance; general superiority over conventional manufacturing; a single universal manufacturability phase transition.

Adversarial conclusions

The conventional joint controller ties exactly. Shared reference bias produces false certification. Independent large seal variation can violate every local object certificate while one global response still passes. These negative or limiting results are part of the main manuscript and are not omitted from the release.

New v3 primary-source checks and novelty boundary

• P15: Prabir Barooah and Joao P. Hespanha, Error Scaling Laws for Linear Optimal Estimation from Relative Measurements, arXiv:0904.2863 (2009). Inspected the primary PDF, including the covariance/effective-resistance identity and dimension laws. These are established prior art, not v3 discoveries. Primary paper.
• P16: Laura Balzano and Robert Nowak, Blind Calibration of Sensor Networks, IPSN 2007. Inspected the author-hosted paper, including the unidentifiable global gain. The v3 absolute-scale ambiguity is consistent with this prior literature. Primary paper.
• P17: Randolph E. Elmquist and coauthors, NIST Measurement Service for DC Standard Resistors, NIST Technical Note 1458, December 2003. Inspected title/authorship and ratio-comparison/reference-maintenance sections. Ratiometric comparison is established measurement practice; it does not eliminate differential errors. NIST note.

The first-party primary documents were used for the specific claims above. The search was scoped, not an exhaustive novelty or patent clearance. No public primary-paper PDF is redistributed as if it were the user's own research.

Candidate contribution: choosing a reachable representative of an explicitly allowed passive-response quotient, exact reserve feasibility/minimum material, its ideal finite-size yield boundary, and a finite-confidence constructive maturation theorem preserving the metrology graph.

Established mathematical ingredients: scale invariance, positive-cone spectral bounds, interval intersection, monotonicity of positive linear costs, order statistics, relative-measurement least squares, effective resistance, sub-Gaussian concentration, local graph relaxation, and gauge ambiguity.

Negative evidence retained: 36 false accepted objects under differential bias; no cycle-residual guarantee against a gradient bias; no reachability at the prescribed fixed representative or low reserve; loss of unfinished repair when the witness link is removed; exact conventional-controller tie.

The original v2 package and its exact bytes are preserved in prior_release/. Current-source validation does not turn source research into independent experimental confirmation.