PDF and manuscript audit
2026-10-01. All 123 PDFs are retained, with a machine-readable catalog. The grouped Papers page is generated from it. No PDF was deleted or rebuilt for this navigation task.
All PDF text extracted successfully. The 115 corresponding TeX bodies match current maintained notes after normalizing Pandoc anchor wrappers, line wrapping and table-width macros. PDF prefixes and TeX headings were screened for gross mismatches; the eight named no-note manuscripts were read in full in both extracted PDF and TeX form. This is document/version screening, not a fresh proof audit or a claim of byte-identical rendering after recompilation.
Classification by role
| Category | Count | Meaning |
|---|---|---|
| current synthesis | 7 | A maintained broad entry point; later live sections still govern frontier details. |
| current technical manuscript | 66 | Current scoped result or obstruction; relevance and proof status remain separate. |
| working manuscript | 30 | Contains useful results while a central comparison, transfer or mechanism remains conditional/open. |
| older version | 9 | Earlier route better framed by a correcting or stronger note; artifact age alone does not determine this label. |
| historical research snapshot | 4 | Source/provenance or superseded programme context, outside the live task queue. |
| orphan manuscript | 7 | No matching notes filename; may retain unique proofs or a valuable standalone exposition. |
Eight manuscripts without a matching note
| Manuscript | Retained value and placement | Current route |
|---|---|---|
| action-gap-foundations | Exact matched-endpoint constant-force chord excess F^2 T^3/(24m), its tangent-area relation, and the zero-infimum distinction between finite variations and their quadratic cost. The conditional finite-copy two-arm discrimination threshold and its collapse as the copy count grows when visibility is below one; its supplied quantum states, phase rule, and measurement model remain assumptions. The explicit separation of Jacobi stability, free-kernel short-time behavior, and relativistic speed-margin examples from any universal action-floor claim. Placement: technical appendix. | action-scale-obstructions |
| classical-action-field | Bounded-speed displacement variance gives a short-window coefficient at most mu^2 Delta and hence zero as the window shrinks. For a finite irreducible reversible velocity chain, the exact positive Green–Kubo plateau is 2m⟨v,(-Q)^{-1}v⟩, with spectral bounds; the two-speed case gives mu^2/lambda. The iterated rate/duration limits differ under the formal scaling u_lambda^2=a lambda, while a fixed speed ceiling prevents this scaling from supplying a physical universal constant. Placement: technical appendix. | composition-universality |
| classical-spins-operational-closure | C125: affine orientation effects give the Bloch ball and capacity two, while the specified classical composites have capacity 2^n and dimension 4^n but product-only pure states, so purification and the stated subspace requirement fail. C127: the reversible H=J u_z v_z interaction evolves identical retained first/cross moments into different product-test probabilities. C128: for all Borel preparations the interaction has no finite-dimensional exact observable closure, including nonlinear updates of finitely many expectation coordinates. Placement: teaching standalone. | classical-orientation-closure |
| collision-action-relaxation | The refreshed elastic bath has exact covariance plateau (m+M)s^2/nu and physical-time exponential relaxation; rescaling the refreshed velocity amplitude makes the plateau arbitrarily small. C050: the equal-mass Poisson hard-point gas gives an infinite-gas plateau mu/rho at collision rate rho u. C051: ordered counterpropagating lattice streams with the same m,u,rho and mean collision rate have periodic tagged motion and zero long-window response; the difference is the joint position–velocity preparation. Placement: technical appendix. | ordered-beam-preparation |
| regulator-limits | For arbitrary partitions, the normalized Gaussian bridge has exact determinant and covariance formulas and its polygonal paths converge uniformly under a common Brownian-bridge coupling as the mesh fluctuation scale vanishes. With (N−1)kappa tending to a positive constant, the paths converge to the straight path while the excess action converges in L2 to half that constant; the finite-dimensional action distribution still has values arbitrarily close to zero. The bounded-speed sine example, fixed-dimensional stationary-phase/delta-weight calculation, and two-regulator parameter map preserve the distinction between a residual limit and a hard finite-path floor. Placement: technical appendix. | two-regulator-audit |
| research-programme | A dated record of the programme map and research priorities as of 2026-09-16, including the provenance and then-current organization of action-selection, operational-closure, and mass-gap routes. Historical context for claims and publications assembled at that date; it contributes no independent theorem and explicitly labels itself superseded by AGENTS.md. Placement: history. | programme evolution |
| same-collisions-different-transport | An accessible derivation of the C050 Poisson-gas plateau and C051 ordered-stream vanishing response under matched mass, speed, density, and mean collision rate. The Palm residual-time distinction: the ordered preparation has a uniform residual wait even though its long-run collision rate is the same; the manuscript explains the full-period cancellation and includes teaching questions. Placement: teaching standalone. | ordered-beam-preparation |
| time-refinement | Exact Gaussian convolution under arbitrary insertion of time points leaves every positive action parameter free, with the deterministic zero-width member also allowed. The bridge mean and variance concentrate on the straight path as the scale vanishes, without selecting the scale; the free unitary Fourier multiplier has generator H_kappa=kappa2(-d2/dx^2)/(2m) on the stated domain. The analytic continuation from the positive Gaussian family to oscillatory unitary evolution is identified as a model choice, not a consequence of refinement. Placement: technical appendix. | refinement-composition-and-limit |
These files are not disposable duplicates. classical-action-field, collision-action-relaxation, regulator-limits and time-refinement retain detailed derivations only summarized or routed by their notes. The spin and matched-collision manuscripts retain coherent teaching/publication presentations of results now split across notes. action-gap-foundations retains its finite-copy operational benchmark. research-programme is a generated historical research document with a source under research/, not an orphan theorem.
The TeX-only spectral-gap-laboratory draft remains outside the accepted PDF set and build integration; the programme/research-note templates are infrastructure, not manuscripts.
Every PDF, classified
Public page changes
Major syntheses appear first, followed by current technical and working manuscripts, then older/historical and unique orphan documents. Every entry links its PDF and current source or correcting treatment and carries an assessment. The page no longer implies all notes have PDFs or all PDFs have equal scientific/publication status. Sensitive titles are qualified by their assessments: Wilson leading versus all-representation activity, the incomplete fixed-lattice Schur proof, preparation-relative record floors, family-dependent Leibniz continuity and the still-open full nonlinear mid-plane comparison.
Regenerate this catalog entry when a PDF or its source materially changes; build only the changed note. The artifact comparison here does not authorize a routine full rebuild.