Changelog¶

All notable changes to this project will be documented in this file.

[Unreleased]¶

Integrate acquired organized-class interaction, effective memory and finite nonlinear comparisons under their declared smooth-sine laws, with shared full-state observation tools and independently retained protocols and evidence. The theory catalog owns the conditional results; no physical constituent identification or unique law follows.
Consolidate advanced-report SDK exports, signed numerical admission and retained trajectory reconstruction. Reject contradictory failure metadata, invalid authoritative aliases and underflow to zero; handle clipped ratios without intermediate overflow and preserve exact values and explicit unavailability.
Refocus the research plan on complete-law admission using organized patterns, with matched target geometry and tangent dynamics. Add a deferred test of law inheritance across scales, retaining hidden state and memory when collective closure fails.
Route structured grammar errors and aggregate validation through the complete shared validator, retaining operator parameters and registry metadata. Unknown operators and unsupported deep recursivity can no longer appear valid through a partial error collector. Trigger documentation checks when the grammar role registry changes, including changes without Markdown edits.
Consolidate the theory catalog and concept cards around maintained owners. Separate native composition, support events, mediated connections and formation proofs into focused chapters while preserving published anchors. Centralize repeated stability/event explanations on their existing contracts and clarify the scope of auxiliary variational models. Preserve nonzero auxiliary gauge-field phases below the former amplitude cutoff, restoring the documented covariant-difference behavior with independent rotation and exact-zero controls.
Share Boolean admission across grammar context, Coupling and Dissonance options, so text such as "false" no longer enables an initialization exception or an optional effect. Preserve Transition's strict Boolean controls across glyphs, public classes and stages. Reject unregistered operators before validating a word. Generate grammar role documentation and operator introspection from their shared registries, and consolidate operator theory around maintained execution contracts instead of retired APIs and untraceable probe figures.
Remove validate_grammar's collect_unified_telemetry argument, which measured an unrelated fixed demonstration graph. Field observations now require an explicitly supplied graph through the existing diagnostic APIs.
Retain the full unordered pair state through phase cancellation in the conservative unit doubled-C5 sine model. Add exact realizability admission, inherited polynomial rates, current derivatives and storage observations with SDK export. The representation needs no midpoint angle, new law or extra continuous degrees of freedom; existing regular adapters keep their own domains.
Consolidate the retained moving-pair interaction contract from existing replica, current and memory results. Distinguish constant collective means from internally modulated compensating currents, and present messages from sufficient future state. Regional-transfer and form-increment reports now attach a rebuilt comparison, preventing stale nested evidence from contradicting the freshly computed ledger.
Connect phase/rate observations to the existing unordered pair state through an exact regular three-node star chart. Recover relative form and inherited evolution, reject singular observation domains, and retain explicit coincident and antipodal counterexamples. This changes representation without selecting a law or removing continuous degrees of freedom.
Consolidate phase information and its motion through shared exact phasor admission, harmonic moments and source/storage calculations. Add a derived phase/rate association observer with an explicit full-star counterexample; no new free parameter or reduced evolution law is introduced. Correct the distinction between odd currents and cancellation under a half turn.
Rebuild sine comparison, mobility and environmental-mediation calculations from admitted primitive state before consuming gradients, rates or work. Altered cached values no longer change those derived results; invalid primitive state remains rejected.
Apply the same report-consumption boundary to joint pairing windows and structural Emission increments. Reuse shared source/window and AL proposal builders so cached bounds, target indices and increments cannot become independent dynamical premises.
Add finite bridge-response discriminators for the sine and one-percent cubic-current laws. Opposite form preparations and an independent phase preparation give exact finite-amplitude curvature coefficients. Shared full-state bounds retain preparation, reading and time-remainder errors; complete endpoint budgets also certify a ratio under a common unknown clock scale within ten percent. These analytical predictions neither install a law nor supply a physical measurement.
Resolve the fixed one-percent cubic-current comparison by a conserved-storage obstruction: the unchanged mapped source cannot acquire either acute unit winding under that law, although its critical geometries remain available. Add a read-only evidence audit without replaying or replacing the frozen sine response; distinguish energetic accessibility from equilibrium stability.
Add a complete smooth phase-storage countermodel with exact local agreement and a different global winding barrier. Integrate a static full-state discriminator and SDK projection, retaining uncertainty, both time directions and the distinction between smooth and real-analytic law classes.
Add a fresh two-direction conservative connection evaluator and freeze its declaration, full response and source. Certify formation and finite acute retention on validated prefixes while preserving the incomplete forward horizon and the separate independent-source obligation.
Reuse cached trigonometric Taylor prefixes without changing recurrence arithmetic; verify nonlinear composition and cache eviction against exact analytic coefficients. Preserve the earlier frozen calculation's source.
Certify metric growth from the actual whole-tube Jacobian using exact rational matrix inequalities and a fixed rate bracket. Reuse signed derivative admission across comparison solvers; retain original-time rates and explicit failures beyond a declared numerical budget.
Add a shared Picard/Taylor step retaining full-state metric uncertainty between steps, with complete local rounding/remainder bounds. Integrate the conservative C5 saddle adapter, original-clock forward/reverse rates, strict whole-time phase admission and SDK export. Keep partial coverage explicit and distinguish solver verification from a formation experiment.
Admit explicit rational near-saddle preparations with retained representation error, resolved near-critical storage and both local directional gates. Centralize full and regional sine storage on the shared 128-bit interval cosine. Add an exact full-state sensitivity metric and conditional nonlinear phase-tube growth bound, without claiming a validated metric forecast or an available zero-winding source.
Prove a same-orbit conservative formation-to-retention construction on C5/private-leaf support, including independent full-state source and target widths and one scaled unit of acute retention. Reuse nonlinear corridors and a phase-band dwell bound in shared analytical reports and SDK exports; keep event-defined centers unavailable and the tiny positive source radius symbolic. Separate this existence result from operational preparation.
Add a full-law conservative saddle-corridor certificate: retained receiver and leaf momentum can select a finite principal-seam passage without a trajectory or tangent extrapolation. Expose the conditional result through shared SDK reports; acute retention and transverse uncertainty remain separate obligations. Extend the cycle-sector obstruction to the exact barrier energy using boundary equilibrium and flow uniqueness.

Conditional dynamics and shared evidence¶

Integrate conservative bridge memory and collective observations with explicit hidden initialization, channel distinctions and source admission. Add form-loss observation obstructions without identifying apparent local attenuation with microscopic dissipation.
Extend phase/storage information and inverse-geometry controls, preserving constitutive freedom and independently frozen response evaluation.
Add retained-frame composition, conservative regional entry/work observers, invariant phase-offset families, finite moving-pattern retention and actual collective-pulse feedback. Formation, maintenance and physical identification retain separate hypotheses; conservative transfer does not select support birth or autonomous preparation.
Preserve positive witnesses, handoff obstructions and frozen negative controls through shared validated flow, exact report projection and SDK export. Correlated cycle storage sharpens target-budget admission; the collective pulse's signed storage split retains its phase-lift dependence.
Add source admission from the conservative C5 sector barrier and a finite rapid-contact averaging bound. The former includes nonflat zero-winding sources; the latter can exclude acute formation despite large storage. Neither passing an admission nor preserving the wider sector proves acute formation, acute retention or physical identity.
Extend the existing conservative entry certificate with full-state source uncertainty and an explicit independently continuable entry box. Separate nominal storage from uncertain source storage and robust short passage from the stronger research requirement of extended finite retention.
Resolve receiver phase velocity, acceleration and third derivative through the existing collective-pulse observer. Retain environmental feedback and document the conditional rigid-motion classification and a compensated initial state whose relative phase motion reappears at third order.
Add a full-state C5 retention window from conserved storage and actual full-support phase-speed bounds. It covers uncertain, internally moving targets without an invariant reference; prepared retention and acquisition from zero winding remain separate claims.
Add a shared reversible-preparation certificate that retains inverse-flow enclosure error and an independent source radius in a strict forward inclusion bound. Reuse the existing frozen producer with a separately admitted declaration mode; preserve historical response schemas and defaults.
Add exact signed-involution reconstruction and a conservative C5/private-leaf saddle assessment. Reuse full support admission and tangent assembly; distinguish invariant families, captured-source membership, local unstable directions and nonlinear passage from actual winding acquisition.

Consolidated documentation¶

Align the introduction, foundations, architecture and public contracts with the implemented mechanisms and their limitations. Connect pulse, resonance, retained memory and scale inheritance without transferring results between incompatible laws or supports.
Keep the execution plan as the sole task queue, the portfolio as branch classification and the strategy as rationale. Frozen evidence and historical source identifiers remain preserved; no new release is implied by this entry.

Cleanup and interface corrections¶

Remove five ineffective global structural-validation settings and normalize the consumed flags/severity through shared parsers. Reject unknown or invalid update batches before committing any setting. Expose that policy as StructuralValidationConfig without changing the existing unified-validator ValidationConfig export, and align runtime and stub declarations. Point the resumption gate directly to its retained pair-state and observation contracts.
Execute the actual README quick start in documentation checks and reject duplicate keys in publication metadata through shared JSON admission. Check finite-work responses against independently constructed doubled-C5 support. Replace two sine-recovery forwarding wrappers with direct shared-owner aliases.
Install declared core dependencies in the publication workflow before running the shared metadata validator; retain one dependency source in package metadata.
Preserve explicitly zero partition overlap in the optional factorization lab instead of silently creating overlapping partitions. Keep test singleton configuration isolated and classical-adapter subtest labels serializable for parallel execution.
Remove unused parallel mathematical score/field helpers, duplicate registry observations and grammar-cache wrappers. Preserve shared nodal calculus, graph fields, live grammar and class lookup; grammar typing now reexports the actual owners, and tnfr.math uses the package version source.
Correct lifecycle phase/coherence formulas, missing-observation status and capacity-clock factors in theory. Retire a duplicate arithmetic campaign driver and unsupported Paley/zero-distance inference; keep scoped algebraic controls and use circular distance in the directed phase diagnostic.
Restore nine advertised tnfr.validation exports through the existing input-validation owner and correct the grammar type-stub import. Retire unused registry no-ops and their counters while preserving live class lookup and registration; API migrations identify the removed interfaces and retained contracts.
Consolidate duplicate equivariance, arithmetic and gauge discussions into their current proof owners. Simplify architecture and workflow descriptions; retain exact recovery records, conditional proofs and negative controls.
Correct the directed-diffusion instrument with an exact weighted growth counterexample. Restrict arithmetic comparisons to their supplied fixtures; spectral multiplicity alone no longer certifies a symmetry group. Incomplete group enumeration and absent counterexample evidence cannot pass the control.

[0.0.3.8] - 2026-09-27 — Relational patterns and consolidated engine contracts¶

Relational dynamics and research¶

Integrate an explicitly selected joint form-phase model with held capacity, shared nodal stepping, SDK access and detached observation reports. Its pressure and storage premises remain distinct from the operator runtime.
Document conditional recovery and formation on supplied support, regional composition and transmission, and reduced descriptions with derived memory. Retained finite responses and mathematical results identify their own scope.
Add read-only attachment and bridge-relocation observations. Report exact differences of represented storage/rates, support cuts, transport reset and explicitly supplied work through shared engine and SDK owners.
Prove a conditional, identity-preserving passive relocation for two prepared five-node rings, including an open recovery region. One supplied event is not an autonomous support law or a repeated-switch stability theorem.
Make support-event choice and its clock the next research gate. A storage budget constrains admissibility but does not force an event or select its time. Physical identification and a uniquely justified full model remain open.

Engine and public interfaces¶

Centralize represented-real admission for pressure, phase, capacity, nodal arithmetic, coherence observations and validation consumers. Retain explicit unavailable evidence and reject invalid inputs before coercion or caching.
Reject NumPy booleans as neighborhood radii across supported NumPy versions, while preserving integer-index inputs for structural and phase observations.
Consolidate JSON input/output and detached report serialization, including duplicate/colliding keys, unsupported labels and non-finite values. Reports and construction recipes remain distinct from resumable checkpoints.
Correct cache ownership, service replacement, operator metric capture, phase proposal validation and shared SDK/CLI diagnostic behavior.
Preserve explicit model, clock, forcing, event and numerical provenance; diagnostic policy labels do not become autonomous dynamics.

Documentation, applications and maintenance¶

Rewrite the README around the current model and research direction, with an accessible summary. Keep AGENTS and its mirror as contributor instructions; consolidate the theory catalog, glossary, contract modules and usage guides.
Move the optional arithmetic projects to applications/primality-test/ and applications/factorization-lab/. Update checkout paths and integrations; these projects are not bundled into the core TNFR wheel or source archive.
Retire obsolete benchmark wrappers, duplicate illustrations and transient application outputs while retaining maintained entry points and scoped evidence. Documentation links and glossary structure have automated integrity checks.
Keep the routine engine/API test gate separate from explicitly selected research suites. Reuse expensive fixtures without weakening independent numerical controls, unsupported-domain checks or execution provenance.

Compatibility¶

Callers relying on permissive numeric/JSON coercion must supply values within the documented domain. Use the shared relational report exporter for exact rational evidence; serializing a report does not execute its proposed event.
Repository scripts using former top-level arithmetic application paths must migrate to applications/. See its index and each application's setup guide.
The 13 named operators and general runtime remain available. Selecting the relational model is explicit; upgrading does not install an autonomous topology selector or establish a physical measurement bridge.

[0.0.3.7] - 2026-09-20 — Nodal engine refactoring and foundation reassessment¶

Foundation reassessment and scope reduction¶

Correct retained phase-response controls to account for represented pressure and endpoint defects instead of assuming exact periodic neutrality. Separate capacity-only ideal profiles from full represented phase forcing.
Separate runtime callback/sampling ordinals from metric retention and physical time. Track admitted invocations, preserve one index across both callback boundaries and reject recursive same-graph execution.
Integrate that ordinal with REMESH cooldown, Recursivity and diagnosis; preserve bounded histories during raw REMESH and detach validation history without duplicating live advisory events.
Execute the predeclared finite P3 phase-to-form comparison through shared pressure/integration owners. Retain independent modal predictions, numerical defects and evidence hashes without selecting a default physical law.
Expose independent availability, provenance and errors for optional substrate, wave-pulse and resonance telemetry without changing their numerical models.
Detect direct NetworkX support/weight/length edits in the shared edge-cache owner, including same-size rewiring that previously retained old pressure/Si inputs. Preserve explicit cache dependency and concurrent-read boundaries.
Unify AL/SHA lifecycle proposals, enforce AL's clipped nondecrease contract before writes and validate SHA/VAL/NUL raw inputs and active log sinks.
Require fresh consecutive valid observations for runtime early stopping, support bounded built-in histories and activate explicit CLI stopping options. Inspect only the configured history tail.
Apply HISTORY_MAXLEN to samples within each metric stream, preserve metric names, resize existing streams consistently and support deque coherence tails.
Detach fluent and provenance reports while preserving opaque node labels; expose unavailable diagnostics and reject invalid study report values/keys.
Reuse unified tetrad summary reductions, retain fit/fallback provenance and expose unavailable optional statistics instead of filtering invalid nodes. Reject nested JSON key collisions before replacing exported files.
Unify solver clock, retained-rate and consumed-phase admission; retire Boolean duration coercion and prevent cached trigonometry from masking invalid state. The unforced row no longer reads unused phase data on the array path.
Share direct/staged Resonance capacity proposals, reject invalid inactive amplification inputs, and fix graphless unidirectional Coupling's partial-write failure. Retain explicit unique outgoing-support semantics and prevent consumed nonzero glyph factors from being coerced into a zero policy.
Admit equivalent reciprocal directed transport in the Euler modal diagnostic through the shared symmetry owner; validate auxiliary reaction rates.
Reject invalid local coherence and colliding node labels in SDK reports; distinguish unavailable logical verdicts from false observations.
Replace obsolete error-reference links and unsupported generic bounds and monotonicity claims with current conditional contracts.
Share validated selector snapshots across execution paths, renew normalizers between decisions and release engine-owned batches on failure. Reject malformed coefficients and normalize large finite mixtures without overflow.
Preserve explicit computation services, resolve registered strategy tokens, and learn from the executed strategy rather than an unresolved AUTO request.
Preserve diagnostic availability in SDK summaries, nodal-rate predictions and phase observations. Stabilize phase imbalance/fits and retain a disconnected-copy control showing why configured phase labels cannot establish emergence.
Reuse nodal admission in pressure hooks and phase/capacity execution; stage phase proposals before writes and share the runtime branch predicate with finite forcing observations.
Recover representable temporal rates from overflowing intermediate differences, reject nonzero underflow rates and detach retained conservation evidence from public snapshots/reports. Admit combined SDK evidence before appending it and compare endpoint energies for nonincrease. Share charge and normalized-energy totals across live fields and retained observations, recovering finite cancellation and exceptional-range quadratic totals.
Share scaled paired correlation between field and sector diagnostics; distinguish unavailable observations from compatibility zero values.
Retire unpaired-correlation, cancelling-stress and unconnected lattice/sieve studies. Reuse circular/Euler arithmetic in retained auxiliary controls and correct example distinctions between phase laws, waves and nodal diffusion.
Share strict represented-real admission across validated nodal rates, equilibrium and sense-index reads; reuse the EPI derivative in the extended model and reject unrepresentable outputs instead of certifying them.
Reuse stable neighbor differences for pressure flux and preserve explicit phase-symmetry availability. Retire the inert-intensity coherence-length sweep.
Expose incremental SDK operator failures, validate candidate lists before execution and remove the unsupported coherence-preservation test claim.
Define one routine engine/API pytest gate through standard testpaths; retain research/certificate suites through explicit paths and make test-all. Remove the duplicate Python 3.11 SDK workflow already covered by main CI.
Remove redundant validation grids and unsupported energy/grammar fixture campaigns. Share expensive arithmetic, phase-ramp and auxiliary-model producers while retaining independent identities, negative controls and rollback boundaries.
Unify circular differences and field reductions across execution paths; retain tiny phase separations and finite inverse-square responses at extreme represented scales, with explicit metric and overflow admission.
Honor the live Gamma registry in scalar and array integration, reject invalid runtime sources and invalidate phase-dependent forcing caches on phase changes.
Reuse stable signed means, dispersion and circular-mean availability in SDK reports; validate authoritative phase values and reject nonfinite JSON export. Correct directed density and share probability admission across SDK recipes.
Unify support and pressure evaluation across dense/multigraph paths and optimization adapters. Preserve finite small pressure under extreme weights and use shared nodal arithmetic instead of reordered products.
Share clipping-policy admission between default/extended integration and held-step observations. Invalid policies reject instead of silently changing the selected projection.
Compute conservation sectors from their captured graph divergences, reject missing or invalid temporal evidence, and distinguish a tolerance alert from an actual non-increasing energy sample.
Make foundational premises and model alternatives explicit; defer the C5 pressure-gate continuation pending minimal-model assessment. Preserve exact conditional results and counterexamples rather than treating previous defaults as uniquely derived laws.
Correct NFR observation availability and configured topology interpretation, separate fitted correlation from spectral fallback, and restrict grammar calibration to its represented monotone domain.
Remove ungrounded chemistry/quantum mappings, external fluid and physical-gap programmes, circular type-signature probes, obsolete examples and their dedicated tests. This intentionally removes public APIs; the retirement record lists affected interfaces, reasons and recoverable source revisions.
Consolidate ontology into a mechanism reuse map and remove duplicated old notebooks from the working tree. The execution plan remains the sole queue.

[0.0.3.6] - 2026-09-19 — Nodal research integration and documentation consolidation¶

Nodal dynamics and research¶

Integrated shared phase/form observations, geometry realization and joint quotient contracts with explicit hypotheses and representation boundaries.
Added finite-state, phase-origin, reflection, hidden-form, support and memory regression cases; retained counterexamples and unavailable-field evidence.
Unified signed scalar EPI admission and clarified continuous integration, prescribed forcing, operator jumps and observed temporal evidence.
Included the post-0.0.3.5 diffusion, finite REMESH/Reception, THOL and grammar work already integrated on main. Its exact and finite-runtime claims retain their specified domains; the current execution plan owns their detailed status.

Documentation and diagnostics¶

Consolidated theory, architecture, developer instructions and the research queue; archived long historical notebooks and superseded proposals unchanged.
Corrected unsupported conservation, complete-state, autonomous-emergence and arithmetic claims in guides, source explanations and release metadata.
Removed the educational manuscript and its site navigation; maintained technical guides and examples own the current explanations.
Generate the operator contract table from the registry and reject metadata drift. Validate local/current-main references, directory fragments and reference-style links; retain checks under optimized Python.
Build one MkDocs site with working mathematical typesetting, diagrams, current guide navigation and historical downloads.

Maintenance and compatibility¶

Added a shared SDK study API and the tnfr network, operators and sequences commands, also available through python -m tnfr. Validated JSON recipes use existing topology builders and operator words; reports retain input state, versions, diagnostic availability and estimator provenance.
Added a runnable study example and one CLI/SDK guide. Recipe cycles count operator words, not physical time; recipes are not complete engine checkpoints.
Keep CLI logs on stderr without replacing application logging handlers. Structured stdout is portable to Windows code pages, and report files use the shared atomic JSON writer. Reject conflicting inputs and invalid step counts before execution. Remove unavailable legacy profiling commands.
Removed dead workflow/profiler entry points and two obsolete benchmarks; current owners and original/archive hashes are recorded in the retirement manifests.
Reduced documentation dependencies to the active MkDocs toolchain and release tooling to build/Twine. Removed unused, contradictory automatic-version settings.
Retained the existing GitHub-release-triggered PyPI publication route. Publication metadata and the release tag must match the package version. Release preparation verifies the selected commit; the workflow itself does not provide a numerical stability certificate or upload to Zenodo.
Scalar-only operations explicitly reject complex/nonuniform EPI payloads. Diagnostic thresholds, fitted coherence length and auxiliary models retain their documented scope; review the current API contracts when upgrading.
Repaired advanced primality CLI validation-schema handling and separated human-readable progress from JSON output. These are interface corrections, not changes to the primality calculations.
Fixed a missing plotting-padding reference, preserved compatibility reexports explicitly and applied the configured Python formatting tools.
Declare SciPy as a core dependency for the SDK's spectral import path. Missing optional system-memory telemetry no longer prevents CPU computation interfaces from loading.
Keep spectral-cache writes from signature readouts off the caller's graph, including snapshot-only calls and graphs with an existing cache.
Handle the empty disagreement sector explicitly in matrix-exponential and transient-gain calculations. Accept UTC Z and variable-width fractional seconds consistently on Python 3.10, preserving raw timestamps and truncating to microseconds without rounding. Retain missing, nonmonotone and mixed-timezone evidence. Compare FFT phase preservation by circular separation at the existing tolerance.
Make CLI byte-lineage tests independent of ignored historical artifacts, remove a pytest-benchmark fixture-name collision and compare spectral estimates against an analytic reference. CI retains per-Python JUnit reports.
Classified the expensive C6 historical report producers as slow tests; their lightweight input-validation and independent exact-model checks remain in the default selection. No producer result is inferred from that classification.
Retired self-contained toy tests, copied verifier criteria, duplicate test runners and obsolete curvature/timing campaigns. Retained their useful contracts in tests that call the actual engine and preserved retirement reasons.
Replaced introductory toy dynamics with the shared SDK entry point. Removed misleading standalone demonstrations and kept explicit model boundaries for retained research examples.
Centralized example loading and CLI checks; retained independent numerical assertions while reducing duplicated cold-import processes. Parallel CI keeps each test file together so module fixtures run once per worker.
Reuse the temporal evaluation's existing window count instead of recomputing its complete field series. Test report wiring separately from numerical work.
Share diagnostic graph preparation while retaining fresh mutable graphs. Remove ineffective test resets for absent subsystems and retain the actual cache, callback, RNG and transaction isolation controls.
Compose testing/development extras from their existing dependency groups. make test follows the configured pytest suite; research producers have separate explicit commands. Benchmark wrapper checks no longer accept arbitrary exceptions as successful tests.

Evidence and open scope¶

Release-specific validation belongs to the GitHub release notes and attached receipts. The scientific objective of deriving autonomous persistent patterns and connecting them to independently specified terrestrial measurements remains open. This release does not establish general C6 stability, a complete state basis, a Millennium conjecture or physical particle emergence.

The detailed pre-release development log is recoverable from Git through the historical record. Descriptions in older release entries below record their original context; current mathematical authority belongs to the theory index.

[0.0.3.5] - 2026-06-24 — Tetrad correspondence audit & emergent redesign¶

A computational audit of the "Universal Tetrahedral Correspondence" found that only π is a genuine structural scale; the four-constant correspondence (φ↔Φ_s, γ↔|∇φ|, e↔ξ_C) is mostly an organizing overlay. Several thresholds asserted as "derived" were empirical, inert (magic), or measured false. This work corrects the claims and replaces magic thresholds with emergent, system-measured quantities. The nodal equation, the 13 operators, and grammar U1–U6 are unchanged.

Corrected (canonicity claims)¶

Only π is a genuine structural scale — the phase-wrap bound shared by BOTH |∇φ| and K_φ (both are means of wrapped angles, ≤ π). γ, e, φ are recoverable as mathematical identities but are NOT the structural scales of their tetrad fields. K_φ matches L_rw·φ only on a smooth, consistent unwrapped branch with matching normalization; the reported correlation is a finite-protocol measurement. The exact wrapped computation is nonlinear; fitted ξ_C uses 1/√λ₂ as a spectral comparison/fallback (not base e).
|∇φ| bound corrected from γ/π ≈ 0.1837 to the phase-wrap bound 0.9π in physics/variational.py, symmetric with K_φ. The measured synchronization onset is ≈ 0.29 and σ-dependent, NOT the constant γ/π; γ/π is retained elsewhere only as a heuristic early-warning level, explicitly labelled non-derived.
derive_tetrad_threshold_values rows re-statused: π geometric; φ, γ, e overlay (recoverable identities, not structural scales).
ARCHITECTURE.md, .zenodo.json, CONTRIBUTING.md, theory/README.md, docs/STRUCTURAL_FIELDS_TETRAD.md, constants/canonical.py — removed the "Universal Tetrahedral Correspondence foundation / 100% derived / zero empirical tuning / verified to machine precision" claims; replaced with the honest tiering (π genuine; γ/e/φ notational overlay).
Second audit pass (repo-wide) — removed the remaining "Universal Tetrahedral Correspondence / canonical derivation / Kuramoto critical coupling" claims and internal contradictions across ARCHITECTURE.md (the "Mathematical Purity / 497 magic numbers eliminated / zero empirical" sections), README.md, theory/FUNDAMENTAL_THEORY.md, theory/GLOSSARY.md, theory/EXTENDED_FIELDS_AND_DERIVED_QUANTITIES.md, AGENTS.md (+ mirror), docs/grammar/PHYSICS_VERIFICATION.md, and the central constants modules (telemetry/constants.py, mathematics/unified_numerical.py, config/defaults_core.py, physics/signatures.py, operators/grammar_telemetry.py, physics/emergent_chemistry.py) plus four benchmarks. Exposed cosmetic "derivations" (e.g. MIN_BUSINESS_SENSE_INDEX = 1/φ + 0.082 ≈ 0.700, ⌊φ×10⌋ = 16) as calibrated/notational values.

Changed (emergent replacement of magic thresholds)¶

physics/phase_transition.py redesigned around a standardized signed-mean ratio. The "universal critical exponent γ_c = γ/π" was measured false (the fitted exponent is protocol-dependent), and the classification noise floor (γ/π)² was proven inert (it sat in a two-order-of-magnitude gap; sweeping it changed no classification). Removed the magic constants GAMMA_C, ORDER_PARAMETER_NOISE_FLOOR, CHIRALITY_THRESHOLD and the theoretical_exponent field; added symmetry_zscore(mean, var, n) = |mean|/√(Var/N) and the selected cut Z_SIGNIFICANCE = 1. The later S6 audit recorded above corrects its interpretation: coupled graph nodes do not justify an independent-sample significance test, so the ratio is an operational spatial-imbalance diagnostic.
Named γ/π constants relabelled as heuristic / non-derived in constants/canonical.py (CRITICAL_EXPONENT, GRAD_PHI_CANONICAL_THRESHOLD, PHASE_GRADIENT_THRESHOLD_CANONICAL) and mathematics/unified_numerical.py; gauge.py, emergent_chemistry.py, interactions.py regime thresholds marked calibrated/heuristic, not derived.

Notes¶

Third audit pass (repo-wide, exhaustive) — removed the remaining "Universal Tetrahedral Correspondence / canonical derivation / zero empirical fitting" claims across the two subprojects (primality-test/, factorization-lab/), examples, benchmarks, the mathematical_purity tests (now check genuine bounds, not the refuted mapping), all theory docs (FUNDAMENTAL_THEORY.md §4 and GLOSSARY renamed to "structural-field tetrad"), and ~40 in-code combo comments (defaults_core.py, bifurcation.py, cycle_detection.py, number_theory.py, etc.) now marked notational. Exposed the primality coefficients (ζ=φγ, η=(γ/φ)π, θ=1/φ) as combos chosen to approximate the original empirical values (ζ=1.0, η=0.8, θ=0.6).
The gauge force-regime classification and emergent-chemistry excitation scale were flagged here for full emergent rework; that rework is now complete — see "Emergent redesign" below.
Full test suite green (2196 passed) after all three passes.

Consequences audit (computational impact, 2026-06-21)¶

A measure-first review of whether calculations (not just narrative) depended on the refuted tetrad values. Main finding: the significant results are robust, because they emerge from STRUCTURE (integer orderings, exact zeros, relative scores) rather than the scale values (γ/π, etc.):

Emergent chemistry (periodic table, magic numbers 2/10/18/36/54/86, octet): the aufbau filling order uses only the integers (n+l, n); the octet is the exact ΔNFR=0 zero — both independent of any scale coefficient. The nu_excitation (=γ/π), nu_0, coherence_gap fields were DEAD (defined, never consumed) and were removed; only theta_valence enters, as a positive scale (the zero is robust to its value).
Number-theory primality (n prime ⟺ ΔNFR=0): each pressure term vanishes individually for primes (Ω−1, τ−2, σ/n−(1+1/n) are all 0), so the result is independent of the coefficients ζ, η, θ.
Gauge interaction regimes: dominant_regime is decided by relative scores, NOT the γ/π threshold; above_threshold (which used γ/π) is metadata.
Riemann ζ-bridge buffer γ/π: a regularisation shift whose exact value is immaterial. K_φ asymptotic exponent α≈2.76: a measured fit, unrelated.

Real consequences corrected: - PHASE_CURVATURE_ABS_THRESHOLD = φ×π ≈ 5.083 was a non-physical K_φ bound (|K_φ| ≤ π by phase wrap, so any check using it was a no-op); it was dead code, corrected to 0.9π ≈ 2.827. - STRUCTURAL_STABILITY_AND_DYNAMICS.md §2.2 still described the old γ_c classification table; updated to the emergent z-score rule. - Removed the dead chemistry scale parameters; fixed a residual MATHEMATICAL_DYNAMICS_BASIS.md |∇φ| claim. Full suite green (2196 passed).

Emergent redesign (gauge regimes + chemistry excitation, 2026-06-21)¶

Completed the full emergent rework of the two studies whose conceptual base was the (now-refuted) four-constant overlay. Both were rebuilt to rest on STRUCTURE alone (measure-first; no value replaced by another magic value):

Gauge interaction-regime classification (physics/gauge.py): removed the three overlay threshold constants REGIME_DOMINANCE_THRESHOLD (1/φ), REGIME_STRONG_THRESHOLD (γ/π, "Kuramoto critical coupling in gauge") and the unused REGIME_SECONDARY_THRESHOLD (γ/(π+γ)). The per-sector above_threshold activity flags now use a single parameter-free criterion: a sector is active when its normalised score exceeds the equipartition share 1/N_REGIMES = 0.25 (the maximum-entropy reference, derived from the number of gauge sectors — the four structural channels of the tetrad). Uniform across all four sectors; no overlay constant. New public symbols N_REGIMES, REGIME_ACTIVITY_SHARE replace the removed thresholds. dominant_regime (relative max of scores) is unchanged — it was already robust. Measured: the criterion always flags the dominant sector and additionally marks genuine co-active secondaries.
Emergent-chemistry valence scale (physics/emergent_chemistry.py): removed the last free scale parameter (theta_valence = 1/φ) and the now-trivial EmergentChemistryParameters dataclass. ΔNFR_chem(Z) is now the integer structural distance of the outer shell to a closed configuration, in natural units (one subshell step = 1) — the exact chemical analogue of primality ΔNFR(n)=0. Noble gases (2,10,18,36,54,86) → ΔNFR=0; halogens/alkali → 1; oxygen → 2; carbon → 4. Magic numbers and the octet are unchanged (they always emerged from the integer (n+l) ordering and the exact zero).
Measured (chemistry): tested whether the (n+l) filling order could emerge from the raw Laplacian spectrum of a concentric multi-shell ("onion") manifold. It does NOT — Madelung ordering reflects electron-electron screening absent from a free graph Laplacian. (n+l) is therefore documented honestly as an integer excitation-count rule (total radial+angular quanta), not a spectral derivation and not a constant correspondence.
Tests updated (test_gauge.py: TestRegimeActivityCriterion, equipartition consistency). Full suite green (2195 passed, 2 skipped).

[0.0.3.4] - 2026-06-17¶

This release consolidates the emergent-geometry program, centralizes the operator/grammar/contract layer onto single canonical sources, opens three new TNFR-native Millennium-problem programs, and refactors the documentation to the current engine state. The 13-operator catalog, grammar U1–U6, and the nodal equation are unchanged; everything below either measures structure the nodal equation already contains or removes duplication. Full suite: 2043 passed, 2 skipped.

Symplectic Substrate (historical entry; current scope corrected)¶

This release introduced an auxiliary ambient symplectic phase space initialized from graph-field snapshots. Later audits established that its exact harmonic-flow identities do not derive the full nodal equation or certify the engine operators.

New module: src/tnfr/physics/symplectic_substrate.py — phase space P = ℝ^{4N} with conjugate pairs (K_φ, J_φ) (geometric) and (Φ_s, J_ΔNFR) (potential); symplectic 2-form ω (antisymmetric, non-degenerate, closed); canonical Poisson brackets; H_sub = ½Σ(K_φ²+J_φ²+Φ_s²+J_ΔNFR²) equal to the structural-energy snapshot exactly after adding the held-fixed gradient background; Liouville div(X_H)=0 for the specified harmonic flow. Each engine operator would require a separate Jacobian pullback test.
Auxiliary model structure tower (each measured to machine precision): Noether charges (time-translation → H_sub; geometric U(1) → E_geo = ½Σ|Ψ|²; potential U(1) → E_pot); the compatible Hermitian / flat-Kähler triple (ω, J, g) with J = −ω — so the i in Ψ = K_φ + i·J_φ is the complex structure the substrate induces; complete integrability (action–angle, Liouville–Arnold); Poincaré–Cartan integral invariants; Marsden–Weinstein symplectic reduction; and the hidden U(2) polarization symmetry whose SU(2) part supplies three conserved Stokes parameters on the per-node Poincaré sphere (classical wave polarization — Stokes 1852 / Poincaré 1892 — not isospin or qubits).
Historical threshold overlay: derive_tetrad_threshold_values reconstructs φ, γ, and e from chosen mathematical identities. They are organizational overlays rather than derived TNFR field scales; π remains the phase primitive.
Consolidated entry point: verify_substrate_geometry(G) bundles all certificates into a SubstrateGeometryReport.
SDK: Network.symplectic_substrate() + SymplecticReport, in TNFR.analyze().
Honest scope: a flat, constant-coefficient linear Kähler backbone — a consolidation of geometry already implied by conservation.py + variational.py; it does not resolve any open program.
Demonstrations: examples/08_emergent_geometry/98, 106, 114.

Emergent Geometry — Structural Diffusion (transport layer)¶

The EPI channel of the canonical ΔNFR is the random-walk graph Laplacian −L_rw·EPI (verified to residual ~1e-16), so the nodal equation is literally a discrete diffusion equation with diffusivity νf. From this single identity the engine measures, in TNFR's own variables, a tower of empirically-established transport phenomena.

New module: src/tnfr/physics/structural_diffusion.py — six transport layers: diffusion/synchronization (Fourier/Fick/Kuramoto), overdamped drift (q̇ = νf·F, Stokes/Einstein mobility — corrects the prior "Newton's second law" reading: the bare first-order nodal equation is overdamped, νf is mobility not inverse mass), discrete standing-wave modes (bounded-manifold Laplacian eigenmodes), structural-stability dispersion relation (σ_k = r − νf·λ_k, the spectral form of U2), random walk + effective resistance (Ohm/Kirchhoff), and structural flow (current, Kirchhoff continuity, Ohm).
Overdamped-projection bridge: the nodal equation is the strong-damping limit of the substrate wave q̈ + γq̇ + Lq = 0 with νf = 1/γ; the γ-dial spans diffusion (γ→∞) to standing waves (γ→0).
Honest scope: the EPI-channel ↔ Laplacian identity is exact; the full ΔNFR is multi-channel; λ_2 is purely topological and does not encode any canonical constant (measured negative result).
Demonstrations: examples/08_emergent_geometry/99, 113, 134, 135.

Operator Contracts & Energy — Centralization and Emergence¶

Canonical contract layer: new src/tnfr/operators/operator_contracts.py — the single source of truth for what each operator does to node state, anchored to the direct _op_* effect (TNFR.pdf §2.2.1). Each OperatorContract records the public English name, the primary_channel (one nodal-equation channel: EPI / νf / θ / ΔNFR), the scale (NODE for twelve operators, NETWORK for the U5 operator REMESH), and a verifiable postcondition. The proactive audit (audit_operator_contracts), the reactive integrity monitor (POSTCONDITIONS), and the introspection metadata now all derive from this spec — eliminating the historical drift where scattered copies disagreed (e.g. AL claiming "positive ΔNFR" though _op_AL only raises EPI; RA checked for EPI increase though it preserves identity; VAL/NUL checked |EPI| though they scale νf).
Public English names: the structural-operator name (Emission, Reception, …) is canonical at the public level; the glyph code (AL, EN, …) is the internal symbol.
Energy/coherence are emergent: the structural candidate energy E = ½Σ(Φ_s²+|∇φ|²+K_φ²+J_φ²+J_ΔNFR²) contains no EPI or νf term (measured: scaling EPI or νf leaves E unchanged). The per-operator Lyapunov role in physics/lyapunov.py is therefore re-derived from the canonical grammar U2 role (config.physics_derivation), not from a hardcoded energy algebra: stabilisers {IL, THOL}, destabilisers {OZ, ZHIR, VAL}, the rest neutral. The form-channel operators (AL, EN, RA, REMESH) leave a same-snapshot evaluation unchanged when all derived fields are held fixed because EPI is absent from the formula. This is a dependency statement, not an isometry or a conservation law after pressure and fields are recomputed.
Dual-lever clarified: the two levers are the two right-hand-side factors of the nodal equation — νf (capacity) and ΔNFR (pressure); operators that write the form EPI (the LHS) sit on neither lever.
Demonstrations: examples/08_emergent_geometry/152, examples/02_physics_regimes/115.

Grammar — Single Canonical Source & Formal-Language Characterization¶

Centralization: the operator-classification sets (generators, closures, stabilizers, destabilizers, transformers, bifurcation triggers/handlers) are derived once in config.physics_derivation and re-exported by operators/grammar_types.py. Every grammar consumer — the U1–U6 validator, the secondary sequence validator, grammar_dynamics, the runtime preconditions, the error factory, and the operator metadata — now reads the single source. Parallel hardcoded copies (including a secondary validator that wrongly listed NUL as a U2 destabilizer) were removed and pinned by tests/operators/test_grammar_canonical_consistency.py.
Canonical grammar spec: new operators/grammar_canon.py materializes the U1–U6 role table, the five-type structural typology, and the canonical glyphic macros (anchored to TNFR.pdf §2.3), with a self-consistency check.
Formal-language thread (characterization, demos only): the flat, non-nested default-depth history projection is represented by 320 reachable states and a 52-state minimal complete DFA. Its numerical growth radius is 10.9560791442; both U2 and U4b reduce the radius. The 944-element syntactic monoid is aperiodic (index 4), so this constructed flat language is star-free / first-order definable. Nested THOL[...] syntax requires a context-free, stack-like model, while runtime U3 and reference-dependent U6 remain outside the DFA. The Parry process is a selected maximum-entropy Markov policy on the dominant component, not a TNFR physical equilibrium or grammar validator.
Demonstrations: examples/08_emergent_geometry/139–152.

Number Theory & the Dual-Lever¶

Prime families as orbits on the zero-pressure set {ΔNFR = 0}; numbers as a coupled network (Ω-graded centrality, primes as the transport periphery); the nodal flow on numbers (primes as equilibria, not attractors); primality as grammatical inertness; numbers as free-monoid words with the dual-lever as the two additive gradings (count Ω → ΔNFR pressure, size log → νf capacity); the capacity arm carries von Mangoldt and the prime-ladder Hamiltonian P14 is the capacity-arm operator — locating the Riemann oscillatory obstruction on the capacity axis the per-node substrate is blind to.
Honest scope: these restate classical multiplicative number theory through the grammar/dual-lever lens; they close no open problem.
Demonstrations: examples/07_number_theory/94–97, 100–102, 116, 146–149.

Millennium Problem Programs (TNFR-native reformulations)¶

Three new programs join Riemann / Navier–Stokes / Yang–Mills. None claims a solution — each carries an explicit honest-scope statement and classified obstruction.

P vs NP (PNP-1) — the nodal equation is a gradient flow, so verifying a configuration's coherence is O(|E|) but synthesizing a globally coherent one by relaxation traps in dissonance basins (measured global-optimum hit rate drops monotonically with problem size on frustrated MAX-CUT). Mirrors P≠NP; Branch B open. theory/TNFR_P_VS_NP_RESEARCH_NOTES.md, examples/09_millennium/109.
Birch–Swinnerton-Dyer (BSD-1) — a_p = p+1−#E(F_p) as structural pressure; the accumulated product reproduces the original 1965 empirical rank separation by brute-force point counting. GL(1)→GL(2) gap open; Branch B. theory/TNFR_BSD_RESEARCH_NOTES.md, examples/09_millennium/110.
Hodge (HC-1, scope corrected under Unreleased) — an auxiliary simplicial complex reproduces the standard finite Hodge decomposition and expected Betti numbers. Canonical |∇φ| and K_φ are node summaries, not oriented edge and face cochains, so no tetrad cochain tower has been derived. The (p,p) and algebraicity gap remains outside this baseline. theory/TNFR_HODGE_RESEARCH_NOTES.md, examples/09_millennium/111.

Documentation, Examples & Repository Hygiene¶

README + core theory docs refactored to the current engine state: corrected a real API note (operators are callable, there is no .apply()), added the emergent-geometry section to theory/FUNDAMENTAL_THEORY.md, rewrote the energy classification in theory/STRUCTURAL_OPERATORS.md / theory/STRUCTURAL_STABILITY_AND_DYNAMICS.md to the emergent/grammar-U2 frame, and updated counts.
Examples reorganized into 10 thematic subfolders (01_foundations … 10_applications), resolving the prior 77–86 numbering collision; each file keeps a stable global number. Foundational examples refactored to the canonical Kuramoto phase-synchrony physics.
Documentation-integrity pass: repaired all dangling example/source/.md links repo-wide, pruned 9 obsolete docs/ files, rebuilt the theory/ hub, and resynced the derived .github/agents/my-agent.md mirror.
Deep repo cleanup: removed foreign GraphQL scratch JSONs, a CUDA debug script, a backup test, an empty dead CLI module, and a stale task tracker; fixed a dangling tnfr-validate console-script entry point in pyproject.toml.
Registry consolidation + lint: the SDK fluent glyph→operator map and the lyapunov operator table now derive from the canonical registries; cleared a small set of dead-code lint findings.
SDK fixes: repaired silent no-ops in auto_optimize and evolve_grammar_aware; added a proactive measured operator-contract fidelity audit (net.audit_operators()).

Research Program Milestones (Yang–Mills Y1–Y5, REMESH-∞ N15, Navier–Stokes N16–N17)¶

The Yang–Mills (Y1–Y5) and REMESH-∞ / Navier–Stokes (N15–N17) program milestones below were developed earlier in the cycle and are part of this release. Each carries an explicit honest-scope statement; none resolves a Clay Millennium Problem.

Y5 — TNFR–Yang–Mills Closure / Obstruction Classification¶

Verdict: BRANCH_B_OBSTRUCTION_CLASSIFIED — Y1–Y4 establish a finite TNFR U(1) structural gauge diagnostic surface, but Clay-strength closure requires a new canonical non-Abelian derivation plus a continuum / thermodynamic lower-bound theorem.
New API: classify_yang_mills_closure() in src/tnfr/yang_mills/closure.py, exported from tnfr.yang_mills with YangMillsClosureReport.
Finite TNFR branch: A_FINITE_U1_DIAGNOSTIC_SURFACE when Y4 reports stable finite positive gaps.
Clay-strength branch: B_REQUIRES_NEW_CANONICAL_NONABELIAN_DERIVATION because Y3 remains OPEN_DERIVABILITY_GAP.
Scope discipline: clay_problem_resolved = False; the obstruction is localized, not removed.
Validation: 4 new tests in tests/physics/test_yang_mills_closure.py cover Branch-B classification, report reuse, sampled collapse handling, and package-root import. Y1–Y5 focused run: 33 passed.
Next target: Y6 / Branch-B derivation search for a TNFR-native non-Abelian connection and non-commuting generator algebra. If no derivation exists without external group labels, the programme should pause at Branch B.

Y4 — TNFR–Yang–Mills Finite Scaling Diagnostic¶

Verdict surface: FINITE_SCALING_EVIDENCE or GAP_COLLAPSE_OBSERVED depending on sampled finite graph families. This is a finite diagnostic only, not a continuum theorem.
New API: run_finite_scaling_study() in src/tnfr/yang_mills/scaling.py, exported from tnfr.yang_mills with FiniteScalingPoint and FiniteScalingReport.
Scaling coordinate (historical name corrected under Unreleased): graph node count n under fixed single-snapshot magnitude ratios ρ_U6 = max_i |Φ_s(i)|/(π/2); this does not assess U6 drift. Grouped reports fit finite log-log slopes of mean gap versus n.
Scope discipline: Y4 runs while YMG-4 remains open. Therefore finite positive scaling evidence cannot be promoted to a Clay-strength Yang–Mills mass-gap claim.
Validation: 6 new tests in tests/physics/test_yang_mills_scaling.py cover report shape/scope, grouped finite scaling, reproducibility, sampled collapse classification, invalid input rejection, and package-root import. Y1–Y4 focused run: 29 passed.
Next target: Y5 closure / obstruction classification, likely Branch B unless a later TNFR-native non-Abelian connection and generator algebra are derived.

Y3 — TNFR–Yang–Mills Non-Abelian Derivability Audit¶

Verdict: OPEN_DERIVABILITY_GAP — audited candidate routes for deriving a non-Abelian / multi-channel gauge sector from TNFR-internal data only; no route is promoted to canonical status.
New API: audit_nonabelian_derivability() in src/tnfr/yang_mills/derivability.py, exported from tnfr.yang_mills with NonAbelianCandidateAudit and NonAbelianDerivabilityReport.
Routes audited: U5 nested-EPI multiplets, THOL/REMESH operator-history internal spaces, and graph cycle-basis bundles.
Obstruction: current canonical Ψ = K_φ + i·J_φ gauge structure supplies a scalar local U(1) connection. Nested EPI or operator-history data do not yet derive component-mixing parallel transport or non-commuting generator algebra; cycle-basis routes require non-canonical basis/orientation selection.
Validation: 5 new tests in tests/physics/test_yang_mills_derivability.py cover baseline U(1) confirmation, nested-EPI obstruction, cycle-bundle rejection, unsupported route errors, and package-root import. Y1+Y2+Y3 focused run: 23 passed.
Open boundary: YMG-4 remains open. Y4 scaling can proceed only as a conditional finite diagnostic; it cannot become a Clay-strength claim while non-Abelian derivability is unresolved.

Y2 — TNFR–Yang–Mills U6 Confinement Sweep¶

Verdict: EMPIRICAL_FINITE_GRAPH_ONLY — finite sweep surface created for testing how the Y1 structural gauge gap behaves across two groups of the single-snapshot potential-magnitude coordinate; the original U6 wording was corrected under Unreleased.
New API: run_u6_confinement_sweep() in src/tnfr/yang_mills/u6_sweep.py, exported from tnfr.yang_mills.
Sweep coordinate: legacy ρ_U6 = max_i |Φ_s(i)|/(π/2); values below one mean below that magnitude scale, not U6-confined. A U6 drift verdict requires a reference snapshot.
Telemetry recorded: gap statistics, self-adjointness, seeded local-U(1) spectral invariance, pure-gauge cycle-closure residuals, applicable snapshot checks, legacy magnitude ratios, and finite-scope metadata. The cycle residual is numerical error, not curvature activity.
Validation: 5 new tests in tests/physics/test_yang_mills_u6_sweep.py cover report shape/scope, U6 target tracking, gap contracts, reproducibility, invalid input rejection, and package-root import. Y1+Y2 focused run: 18 passed.
Open boundary: Y2 does not prove a U6 lower-bound theorem and does not address non-Abelian derivability (YMG-4) or continuum scaling (YMG-5). Next target: Y3 derivability audit.

Y1 — TNFR–Yang–Mills Finite Structural Gauge Gap Diagnostic¶

Verdict: DIAGNOSTIC_SURFACE_CREATED — first TNFR-native Yang–Mills / structural mass-gap attack surface implemented as a finite-graph diagnostic, not a Clay-strength proof.
New package: src/tnfr/yang_mills/ with build_structural_gauge_graph(), build_structural_gauge_gap_operator(), and compute_structural_gauge_gap().
Operator (current notation): H_structural = L_A + V_F + V_Phi, where A=d(arg Psi) makes L_A unitarily equivalent to the ordinary weighted graph Laplacian, V_F is analytically zero apart from numerical cycle-closure residuals, and V_Phi is the selected single-snapshot magnitude penalty Phi_s²/(pi/2)²; it is not a U6 test.
TNFR scope discipline: the reported number is the spectral gap of this selected finite matrix. It is not evidence of dynamical gauge curvature, confinement, a physical mass, or a continuum Yang–Mills mass gap.
Validation: 13 new tests in tests/physics/test_yang_mills_structural_gap.py cover graph construction, self-adjointness, non-negative finite gap reporting, seeded local-U(1) spectral invariance, reproducibility, package imports, and no EPI/phase mutation. Focused run: 13 passed.
Open boundaries: non-Abelian derivability (YMG-4) and continuum / thermodynamic scaling (YMG-5) remain open.
Documentation: theory/TNFR_YANG_MILLS_RESEARCH_NOTES.md records the Y-series gap ledger and updates the next target to Y2 (U6 confinement sweep).

N17-A — U3+U5 → K41: Analytical Cascade Locality (ANALYTICAL_CONSISTENT_CONDITIONAL)¶

Verdict: ANALYTICAL_CONSISTENT_CONDITIONAL — K41 \(k^{-5/3}\) spectrum derived conditionally from TNFR grammar rules U2+U3+U5+CDC; algebraically closed given the Cascade Development Condition.
Lemma U5-SS (U5 + U2 → scale self-similarity): U5-uniformity (same canonical operators and constants at every hierarchy level) + U2 force \(u_\ell = C(\varepsilon r_\ell)^{1/3}\) in the inertial range. The K41 scaling emerges from grammar structure (U5 collapses the dimensionless ratio to a level-independent constant), not from external dimensional analysis.
Lemma U3-CL (U3 → cascade locality, conditional): Under Lemma U5-SS, U3 (phase-gated coupling, \(|\phi_i - \phi_j| \le \Delta\phi_{\max}\)) blocks all inter-level interactions if and only if the Cascade Development Condition (CDC) holds → constant energy flux \(\Pi_\ell = u_\ell^3 / r_\ell = \varepsilon\) across scales.
Theorem (U2 + U3 + U5 + CDC → K41): \(E(k_\ell) \sim \varepsilon^{2/3} k_\ell^{-5/3}\) — proof: \(E_\ell \sim u_\ell^2 \sim \varepsilon^{2/3} r_\ell^{2/3}\), \(E(k_\ell) = E_\ell / \Delta k_\ell\) with \(\Delta k_\ell \sim k_\ell\) (log bands) → \(E(k_\ell) \sim \varepsilon^{2/3} k_\ell^{-5/3}\). □
CDC (irreducible gap): CDC (adjacent cascade levels have \(|\phi_{\ell,i} - \phi_{\ell+1,j}| \ge \Delta\phi_{\max}\) for all \(i,j\)) is not derivable from U3, U5, or the nodal equation. It is the K41 locality hypothesis restated in TNFR language, and the structural analogue of \(S(T) = (1/\pi)\arg\zeta(\tfrac12 + iT)\) in the Riemann programme — reachable only by a sufficiently developed turbulent cascade, not from the canonical operator catalog alone.
N17-A does not close NS-G1..G4 — those gaps concern continuum-limit, uniform bounds, BKM criterion, and vortex stretching; not cascade locality.
N17-B pre-registered (deferred): empirical energy spectrum via energy_spectrum_3d() (to be implemented in src/tnfr/navier_stokes/operator.py), n ∈ {32, 48}, ν ∈ {0.01, 0.005}, T = 2.0. Expected verdict: STEEPER_THAN_K41 (CDC not satisfied at Re_eff ≤ 500).
Documentation: theory/TNFR_NAVIER_STOKES_RESEARCH_NOTES.md §20 (full lemmas, theorem, CDC gap analysis, verdict table, N17-B pre-registration spec).

N16 — NS-G5 Closure: 2D-Embedding Lemma¶

Verdict: NS-G5 CLOSED at the discrete-operator level via the 2D-Embedding Lemma (Theorem NS-G5-TNFR).
Algebraic proof (three steps using existing TNFRNavierStokesOperator methods on z-independent u = (u₀(x,y), u₁(x,y), 0)):
vorticity_3d: ω₀ = ω₁ = 0, ω₂ = ∂_x(v) − ∂_y(u)
vortex_stretching_field: S_a = ω₀·∂_x(u_a) + ω₁·∂_y(u_a) + ω₂·∂_z(u_a) = ω₂·0 = 0 for all a
stretching_production: = 0.0 exactly in IEEE 754
TNFR reading: z-channel decoupling → no cross-channel ΔNFR → enstrophy ≤ viscous dissipation (monotonically non-increasing) → discrete TNFR analogue of 2D NS global regularity.
Contrast with 3D: ∂_z(u_a) ≠ 0 activates cross-channel ΔNFR coupling → stretching production generically positive → U2 (convergence/boundedness) is not guaranteed → vortex stretching amplification is structurally active.
Empirical corroborator: examples/85_navier_stokes_dimensional_asymmetry.py — z-independence → stretching_production ≈ 0 at machine precision across all tested configurations (commit 1fac358b).
Scope: NS-G5 closure does NOT affect NS-G1..G4 and does NOT address the Clay Millennium Problem (3D global regularity).
Documentation: theory/TNFR_NAVIER_STOKES_RESEARCH_NOTES.md §19.

N15 REMESH fixed-delay surrogate — corrected historical record¶

Historical milestones: commits a1f298fd, badac156, and 48b0574a introduced a history-space projection, projected charge/energy claims, and spectral comparisons under the label “REMESH-∞ closure.” The corrected analysis is maintained in theory/REMESH_INFINITY_DERIVATION.md §§1–23.
Valid restricted result: on a finite cyclic history window, the unclipped fixed-coefficient filter \(F=\beta I+\gamma S^{\tau_l}+\delta S^{\tau_g}\) is a normal contraction for \(0<\alpha<1\). Its Cesàro averages converge to the orthogonal projection onto \(\ker(I-F)\). The common fixed modes have period \(\gcd(\tau_l,\tau_g)\); the former lcm lattice statement was incorrect.
Runtime boundary: apply_network_remesh is history-gated, configurable, structurally clipped, and uses a snapshot whose location changes with \(\tau_g\). The finite cyclic result does not establish its literal \(\tau_g\to\infty\) limit.
Invariant correction: convex EPI mixing is not an isometry. Orthogonal projection contracts the surrogate history norm, but it neither proves conservation of TNFR structural charge nor monotonicity of the structural candidate energy. The former exact-conservation and universal \(O(1/n)\) rate claims are superseded.
Registry scope: the surrogate projection can be computed without adding an implementation registry entry. This does not prove that the 13 registered operators exhaust all admissible TNFR transformations; catalog completeness remains open.
External-program scope: the finite surrogate does not advance RH, Navier–Stokes regularity, K41, or an RMT correspondence. Historical branch labels now apply only to comparisons made inside that declared surrogate.

[0.0.3.3] - 2026-03-07¶

Documentation Audit (Sessions 1-4)¶

Comprehensive tone audit: Removed speculative/grandiose language across 25+ files
TNFR_RIEMANN_RESEARCH_NOTES.md: Reduced from 2679 to 1499 lines (removed unfounded claims)
AGENTS.md: Fixed 'inevitability' → 'derivation strength', updated conservation test count (62 → 88), verified all 40+ cross-reference links
Synced .github/agents/my-agent.md with AGENTS.md corrections
Updated test counts to 1,655 across 8 files
Removed orphaned file: src/train_gmx_optimizer.py
Fixed contradictions between AGENTS.md and theory/ documents
Validated: 1653 passed, 2 skipped

[0.0.3.2] - 2026-03-06¶

Documentation & Consistency Fixes¶

Corrected false Γ(4/3)/Γ(1/3) derivation in MINIMAL_STRUCTURAL_DEGREES.md and FUNDAMENTAL_THEORY.md (Γ(4/3)/Γ(1/3) = 1/3, not 0.7711)
Synchronized .github/agents/my-agent.md with AGENTS.md (K_φ threshold, MIN_BUSINESS_COHERENCE, THOL_MIN values)
Fixed CHANGELOG version to match pyproject.toml (0.0.3.2)
Fixed MIN_BUSINESS_COHERENCE precision in ARCHITECTURE.md and CONTRIBUTING.md (0.751 → 0.7506)
Resolved phantom docs/TNFR_FORCES_EMERGENCE.md references across 8+ files
Removed dead code src/tnfr/config.py (shadowed by config/ package)
Cleaned unused imports in sdk/simple.py

[0.0.3] - 2026-03-05¶

Structural Conservation Theorem¶

conservation.py: Introduced a finite-trajectory structural-balance residual and Noether-like vocabulary. Later audits established that U1-U6 do not imply a vanishing residual or a universal conservation theorem.
Added charge-density/current, structural-energy, Ward-residual, stability-policy, and spectral diagnostics; these names do not by themselves establish conserved charges, Noether hypotheses, or Lyapunov monotonicity.
Two-sector structure: Potential (Φ_s ↔ J_ΔNFR) and Geometric (K_φ ↔ J_φ) coupled through Ψ = K_φ + i·J_φ
62 validation tests, charge drift < 0.03% across topologies

Dissipative Conservation¶

dissipative_conservation.py: GPU-accelerated dissipative conservation analysis with PyTorch backend
Phase field computation, dissipation rate tracking, and energy budget monitoring

Closed-Loop Integrity Monitor¶

integrity.py: StructuralIntegrityMonitor with complete postconditions for all 13 canonical operators
Each operator (AL, EN, IL, OZ, UM, RA, SHA, VAL, NUL, THOL, ZHIR, NAV, REMESH) has verified pre/postcondition contracts
Automatic violation detection and reporting

Grammar-Aware Dynamics¶

grammar_dynamics.py: Bridge between grammar validation (U1-U6) and dynamic operator selection
Incremental U1-U6 checks: validate_candidate(), filter_candidates(), suggest_alternative(), enforce_grammar_on_glyph()
Priority-based operator substitution with fallback logic
grammar_application.py: Pre-validation in apply_glyph_with_grammar() for grammar enforcement before operator application
selectors.py: _soft_grammar_prefilter() wired with grammar_dynamics for operator filtering

Simple SDK — Research-Grade Access¶

simple.py: Upgraded with full Structural Field Tetrad, conservation laws, and unified telemetry access
TetradSnapshot dataclass: phi_s, grad_phi, k_phi, xi_c, j_phi, j_dnfr with is_safe() and summary()
ConservationReport dataclass: noether_charge, energy, lyapunov_stable, lyapunov_derivative, conservation_quality with summary()
10 new Network methods: tetrad(), fields(), conservation(), telemetry(), tensor_invariants(), emergent_fields(), evolve_grammar_aware(), integrity_check(), upgraded results() and info()
TNFR.analyze(): One-shot comprehensive analysis (coherence, tetrad, conservation, tensor invariants, emergent fields, integrity)
Feature-gated imports: _HAS_FIELDS, _HAS_CONSERVATION, _HAS_INTEGRITY, _HAS_GRAMMAR_DYNAMICS
29 new tests in tests/sdk/test_simple_advanced.py

Shared Test Infrastructure¶

tests/conftest.py: Centralized test fixtures (make_ring_graph, make_node_data, ring3, ring5, small_graph)
DRY reduction across 16+ test files that previously duplicated _make_graph helpers

Code Quality¶

Fixed bare except: clauses in grammar_dynamics.py (now except Exception:)
NAV bypass fix for grammar validation edge case
Redundancy elimination across physics helpers
Rich operator postconditions (13/13 coverage)

Cross-Codebase Constant Unification (Round 1)¶

grammar_types.py: Eliminated duplicate operator sets (single canonical definition)
THOL_MIN_COLLECTIVE_COHERENCE: Unified to canonical 0.2413 (was 0.3)
MIN_BUSINESS_COHERENCE: Centralized to canonical formula (e×φ)/(π+e) ≈ 0.7506
health_analyzer.py / self_organization.py: Aligned fallback values to canonical

Phase Gradient Threshold Unification¶

Canonical value: γ/π ≈ 0.1837 (Kuramoto critical coupling in TNFR units)
Unified across 9 code files: Replaced competing values (0.2904, 0.2886, 0.2915, 0.38) with single canonical derivation
Updated 8 documentation files: Consistent threshold references throughout

Cross-Codebase Constant Unification (Round 2)¶

compute_structural_potential_field: Added alias in physics/fields.py (was silently missing, imported in 2 files)
SHA_VF_FACTOR comment: Fixed from ≈ 0.8476 to correct ≈ 0.9015 in defaults_core.py
Operator fallback values: SHA (0.85→0.9015), NUL (0.85→0.9015), VAL (1.05→1.0676) aligned to canonical
K_φ hotspot formula: Fixed in conservation.py from 2π/√5 ≈ 2.8099 to canonical 0.9×π ≈ 2.8274
grammar_core.py K_φ default: Fixed from 3.0 to canonical 2.8274
telemetry/constants.py: Removed dead try/except ImportError fallback; direct canonical imports
config.py: Structural field thresholds now derive from constants.canonical (was hardcoded)
pyproject.toml: Added mpmath to core dependencies (was required but unlisted)
Documentation sync: Updated 7 doc files with correct threshold values and test counts

Test Suite¶

1,655 tests (1,646 passing, 9 skipped), 0 failing
Coverage spans operators, physics, dynamics, grammar, conservation, integrity, SDK, and factorization

[0.0.2] - 2025-11-29¶

TNFR Development Doctrine Establishment¶

Foundational Principle: Added TNFR Development Doctrine as core methodological commitment
Theoretical Integrity: Commitment to follow mathematics objectively from nodal equation ∂EPI/∂t = νf · ΔNFR(t)
Scientific Independence: Defend conclusions emerging rigorously from TNFR principles regardless of external paradigm alignment
Validation Criteria: Established 4-point validation framework (Derivable, Testable, Reproducible, Coherent)

Complete Framework Expansion¶

29 New Examples: Comprehensive examples (11-39) covering physics, biology, cosmology, consciousness studies
TNFR-Riemann Program: Complete theoretical framework connecting discrete operators to Riemann Hypothesis
Advanced Physics Modules: Classical mechanics, quantum mechanics, symplectic integration implementations
Extensive Theory Documentation: 25+ specialized theoretical documents in theory/ directory

Documentation Academic Modernization¶

Unified Academic Tone: Systematic elimination of grandilocuent language across all documentation
README Gateway: Transformed main README into coherent documentation entry point
Consistent Terminology: Standardized "Primary theoretical reference" replacing "SINGLE SOURCE OF TRUTH"
Professional Presentation: Enhanced credibility through formal academic language standards

Test Suite Optimization¶

Major Cleanup: Removed 58 obsolete test files (82 → ~30 files)
100% Pass Rate: Achieved 173 passing, 7 skipped, 0 failing tests
Focused Validation: Retained only tests validating TNFR theoretical foundations
Core Coverage: Mathematics, operators, physics, validation maintained

Technical Enhancements¶

Enhanced N-body Dynamics: Improved TNFR integration with classical mechanics
Riemann Operator: Complete implementation with eigenvalue analysis capabilities
Type System: Enhanced type definitions and structural validation
Code Quality: Significant cleanup removing outdated components

[9.7.0] - 2025-11-29¶

Major Theoretical Enhancements¶

Universal Tetrahedral Correspondence: Complete mathematical framework establishing exact mapping between four universal constants (φ, γ, π, e) and four structural fields (Φ_s, |∇φ|, K_φ, ξ_C) (later superseded — see the φ/γ/e purge under [Unreleased]: only π is a genuine structural scale)
Unified Field Framework: Mathematical unification discovering complex geometric field Ψ = K_φ + i·J_φ with emergent invariants
Self-Optimizing Engine: Self-optimization capabilities with unified field telemetry for automated structural optimization
Complete Academic Documentation: Comprehensive conversion to formal academic tone across entire documentation ecosystem

Canonical Invariants Optimization¶

Consolidated from 10 to 6 canonical invariants based on mathematical derivation from nodal equation
Optimized invariants: Nodal Equation Integrity, Phase-Coherent Coupling, Multi-Scale Fractality, Grammar Compliance, Structural Metrology, Reproducible Dynamics
Enhanced theoretical consistency and reduced redundancy

Documentation Modernization¶

AGENTS.md: Complete academic conversion maintaining single source of truth status
README.md: Restructured with new Getting Started section and clear learning paths
GLOSSARY.md: Comprehensive expansion with Universal Tetrahedral Correspondence coverage
Eliminated promotional language and emojis across entire ecosystem
Updated all version references to 9.7.0

Structural Field Tetrad¶

Complete Mathematical Foundations: All four canonical fields now have rigorous mathematical derivations
CANONICAL Status: Φ_s, |∇φ|, K_φ, ξ_C all promoted to canonical status with theoretical validation
Unified Complex Geometry: Integration of curvature and transport via complex field Ψ

Development Infrastructure¶

Updated pyproject.toml to v9.7.0 with current dependency structure
Modernized CONTRIBUTING.md with academic tone and current 6 invariants
Enhanced TESTING.md with updated invariant validation framework
Complete English-only policy implementation

[9.1.0] - 2025-11-14¶

Added¶

Phase 3 structural instrumentation:
run_structural_validation aggregator (grammar U1-U3 + field thresholds Φ_s, |∇φ|, K_φ, ξ_C, optional ΔΦ_s drift).
compute_structural_health with risk levels and recommendations.
TelemetryEmitter integration example (examples/structural_health_demo.py).
Performance guardrails: PerformanceRegistry, perf_guard, compare_overhead.
CLI: scripts/structural_health_report.py (on-demand health summaries).
Docs: README Phase 3 section, CONTRIBUTING instrumentation notes, docs/STRUCTURAL_HEALTH.md.
Glyph-aware grammar error factory (operator glyph → canonical name mapping).

Tests¶

Added unit tests for validation, health, grammar error factory, telemetry emitter, performance guardrails.

Performance¶

Validation instrumentation overhead ~5.8% (moderate workload) below 8% guardrail.

Internal¶

Optional perf_registry parameter in run_structural_validation (read-only timing).
Canonical operator registry frozen (removed dynamic auto-registration, cache invalidation, metaclass telemetry, reload script). Attempting dynamic registration now raises. Ensures strict adherence to unified grammar (U1-U4) and prevents non-canonical transformations.

Deferred¶

U4 bifurcation validation excluded pending dedicated handler reintroduction.

Integrity¶

All changes preserve TNFR canonical invariants (no EPI mutation; phase verification intact; read-only telemetry/validation).
Registry immutability strengthens invariants #1 (EPI only via operators), #4 (operator closure) and #5 (phase verification untouched). Tests updated: removed dynamic registration tests; added test_canonical_operator_set.

[9.0.2]¶

Previous release (see repository history) with foundational operators, unified grammar, metrics, and canonical field tetrad.