# Soliton Bus Model This page documents the finite transport and operator contracts in `Signals.SolitonBus`. The source chats describe a proposed N-LIG/Lignolux architecture. ## Chat-Derived Design The reviewed chat corpus groups the bus design into four related layers: | Layer | Chat-derived proposal | Model boundary | | --- | --- | --- | | Multiplexing | WDM, MDM, OAM, and A2Q soliton-bus addressing | Lane addresses and selectors are finite data; they do not prove independent physical channels. | | Routing | Thermo-optic address trees, demultiplexers, bucket-brigade routing, and OAM sorting | `RoutingPlan` requires an injective destination map; switching latency and crosstalk remain calibration data. | | Readout | Kerr cross-phase modulation, probe-beam phase shifts, and homodyne detection | The phase relation is a conditional interaction contract; it is not a proof of nondestructive quantum measurement. | | Operators | Gaussian quadrature operations, beam splitters, phase gates, CNOT-like gates, and nonlinear phase interactions | Finite matrices and phase-space maps are implemented; material coupling and universal hardware operation are not inferred. | The source request explicitly groups WDM, MDM, OAM, A2Q, squeezed-state readout, and nondestructive phase measurement [in the signal-phase chat](https://github.com/westurner/sustainablefactory/blob/main/data/chats/IQ-Sampling-for-Signal-Phase.md#L8547). The same chat proposes OAM encoding and a Kerr probe path [at the bus design section](https://github.com/westurner/sustainablefactory/blob/main/data/chats/IQ-Sampling-for-Signal-Phase.md#L8589). The proposed N-LIG architecture also describes the signal bus, OAM qudit payload, and a Kerr cavity [in the interaction-zone design](https://github.com/westurner/sustainablefactory/blob/main/data/chats/IQ-Sampling-for-Signal-Phase.md#L8729). The separate architecture chat proposes wavelength as a coarse address and a thermo-optic layer as a fine decoder [in its address-tree description](https://github.com/westurner/sustainablefactory/blob/main/data/chats/_Quantum%20Processor%20and%20Soliton%20Discussions%20%20.md#L363). It describes bucket-brigade routing as a simulation target, with switch latency and loss as quantities to measure [in the network-routing plan](https://github.com/westurner/sustainablefactory/blob/main/data/chats/_Quantum%20Processor%20and%20Soliton%20Discussions%20%20.md#L582). ## Classical Transport API `LaneAddress` represents five independent multiplexing coordinates: - wavelength channel for WDM; - spatial mode for MDM; - integer orbital-angular-momentum charge; - polarization channel; - time slot for TDM. `LaneSelector` enables any subset of those coordinates. `demultiplex` filters finite frame lists by the enabled selector fields. `BusNetwork` and `RoutingPlan` represent a finite lane fabric; the route map must be injective, so the contract does not silently merge two source lanes. `broadcastLanes` represents a finite bucket-brigade fan-out. The classical logic inventory includes identity, NOT, NAND, AND, OR, XOR, XNOR, and NOR. The lemmas `nand_complete_not`, `nand_complete_and`, and `nand_complete_or` show the usual NAND-derived identities for Boolean values. The inventory is complete as a finite Boolean operator set; it is not a throughput or hardware-timing claim. ## Quantum and CV Operators The module uses explicit finite complex matrices and amplitude-vector application, so it does not require a second Lean package or silently import a different mathlib build. The finite one-qubit inventory contains: - Pauli X, Y, and Z; - Hadamard, also exposed as the balanced dual-rail beam-splitter matrix; - parameterized unit-modulus phase gates. The multi-mode inventory contains explicit CNOT, SWAP, controlled-phase, and Toffoli matrices. Standard one-qubit operators carry checked finite unitarity lemmas. Multi-qubit matrices remain explicit operator data until their permutation-unitarity lemmas are added. The finite quadrature layer supplies displacement, phase rotation, reciprocal squeezing, balanced beam splitting, Kerr phase shift, and cross-phase modulation. These are finite mathematical operators, not claims that N-LIG provides the required nonlinear coefficient or low-loss quantum regime. The chat proposal describes Gaussian rotations and squeezing together with a Kerr/non-Gaussian path in `data/chats/Breakthrough-in-Extreme-Dielectric-Nanolasers.md` and proposes cross-phase modulation between OAM solitons in that same chat. Those claims remain conditional in the model. ## Dynamics Contracts The propagation layer now records the standard finite nonlinear-envelope quantities used in hollow-core soliton work: $$ L_D = \frac{T_0^2}{|\beta_2|}, \qquad L_{NL} = \frac{1}{\gamma P_0}, \qquad N^2 = \frac{L_D}{L_{NL}}. $$ `SolitonPropagationDynamics` stores the pulse duration $T_0$, peak power $P_0$, anomalous group-velocity dispersion $\beta_2$, nonlinear coefficient $\gamma$, both characteristic lengths, and the supplied soliton order. The record also keeps self-compression and resonant-dispersive-wave observations as hypotheses, because those outcomes depend on gas pressure, core geometry, loss, mode matching, and photoionization. `SolitonCollisionDynamics` makes the two-pulse interaction explicit: $$ \Delta\phi_{probe,XPM} = \gamma_{XPM} P_{signal} L_{eff}. $$ The phase law is checked algebraically. The coefficient includes the chosen XPM convention and modal-overlap factor; it is not a universal material constant. Collision observation, nonabsorption, and crosstalk remain measured or conditional fields. This is the appropriate boundary for XPM and Kerr proposals: a formula for phase shift is not a demonstrated single-photon gate. This is also a review correction: the phase shift is driven by the interacting signal power, not by the probe power being phase-shifted. Both powers remain in the record so self-phase, probe loading, and collision bookkeeping can be extended without changing the meaning of the XPM law. `OAMMultiplexingObservation` records fibre length, mode-group count, symbol rate, data rate, modal crosstalk, and MIMO equalizer order. The long-haul OAM experiment below demonstrates modal transport and equalization, but it does not demonstrate soliton propagation or N-LIG fabrication. `SolitonDynamicsOperator.defaultPlacement` separates operator location: | Placement | Operators | Interpretation | | --- | --- | --- | | `onBus` | linear propagation, self-phase modulation, XPM, self-compression, OAM mode coupling | Co-propagating or in-guide dynamics, subject to dispersion, loss, and modal calibration. | | `busInterface` | resonant dispersive-wave generation, parity measurement | Requires a spectral or measurement boundary; not a transparent bus primitive by default. | | `offBus` | homodyne readout | A detector/local oscillator path outside the transported signal mode. | ## Multi-Lane ($N > 0$) Interferometric Couplers, $Gr_{\mathbb{C}}(2,4)$ Twistor Channels, and Yang–Baxter Soliton Collisions In [src/signals/Signals/SolitonBus.lean](https://github.com/westurner/sustainablefactory/blob/main/src/signals/Signals/SolitonBus.lean), `SolitonMultiLaneCoupler` connects the $N$-lane soliton-bus architecture to the multi-model interferometric and Quantum Inverse Scattering frameworks developed in [grassmannian-huygens-double-slit-oam-paper.md](grassmannian-huygens-double-slit-oam-paper.md) and [grassmannian-qism-bethe-ansatz-paper.md](grassmannian-qism-bethe-ansatz-paper.md): 1. **$N$-Lane ($N > 0$) Grating Coupler & Barycentric Power Imbalance (`SolitonMultiLaneCoupler`)**: Parameterized by lane count `numLanes > 0`, lane aperture width $a > 0$, lane height $b > 0$, center-to-center waveguide pitch $d > 0$, wavelength $\lambda > 0$, coupler distance $L > 0$, barycentric lane-imbalance shift $d_{\mathrm{bary}} \in [-1, 1]$, visibility $\mathcal{V} \in [0, 1]$, and differential OAM topological charge $\Delta\ell \in \mathbb{Z}$: - For $N = 1$ lane (`SolitonMultiLaneCoupler.singleLane_intensity`), the coupler output reduces identically to the single-waveguide envelope $I_0 E_{\mathrm{slit}}(x, y)$. - For $N = 2$ lanes (`SolitonMultiLaneCoupler.dualLane_intensity_eq_huygensSteinerOAM`), the output matches `huygensSteinerOAMDoubleSlit` identically. 2. **Complex $Gr_{\mathbb{C}}(2,4)$ Twistor Coupler Representation (`SolitonMultiLaneCoupler.dualLane_complexGr24_eq_huygensSteinerOAM`)**: Encoding a dual-rail OAM soliton coupler in the complex Grassmannian $Gr_{\mathbb{C}}(2,4)$ with $\mathbb{C}^4$-projected twistor entries $$ z_{1,2} = \sqrt{1 \pm d_{\mathrm{bary}}}\,e^{\pm i\left(\frac{\pi d x}{\lambda L} + \frac{\Delta\ell\,\phi}{2}\right)} $$ reproduces the partially coherent Huygens–Steiner OAM coupler output with visibility $\mathcal{V} = \sqrt{1 - d_{\mathrm{bary}}^2}$ identically across all $(x, y, \phi)$ with **zero relative predictive error**, eliminating the $50\%$ error of unphased $Gr_{\ge 0}(2,4)$ crossing-minor proxies. 3. **Three-Soliton Yang–Baxter Collision Order Independence (`solitonThreeWaveguide_yang_baxter_consistency`)**: When three soliton lanes with spectral rapidities $(u, v, w)$ and coupling $\eta$ interact pairwise in an integrable NLSE / Yangian zone, `solitonThreeWaveguide_yang_baxter_consistency` combines `Signals.QISM.yang_baxter_equation_matrix8` and `Signals.QISM.baxterTQResidual_wronskian_zero` to certify that the $8 \times 8$ three-mode scattering product satisfies $$ R_{12}(u-v;\eta)\,R_{13}(u-w;\eta)\,R_{23}(v-w;\eta) = R_{23}(v-w;\eta)\,R_{13}(u-w;\eta)\,R_{12}(u-v;\eta) $$ and the three-point Casorati Wronskians on $Gr(2,4)$ satisfy Baxter's $TQ$ difference equation identically. 4. **Rust + WebAssembly (`wasm32-unknown-unknown`) + WebGL2 GLSL Implementation**: All seven multi-slit / multi-lane ($N > 0$) models (`Fraunhofer`, `HuygensSteiner`, `Gr24RealPlucker`, `Gr24ComplexTwistor`, `PilotWave`, `Wigner`, `Gr24Unphased`) are implemented in [src/signals/numerical_adapter/src/double_slit.rs](https://github.com/westurner/sustainablefactory/blob/main/src/signals/numerical_adapter/src/double_slit.rs) with WebGL2 (`#version 300 es`) shaders in [double_slit.vert.glsl](https://github.com/westurner/sustainablefactory/blob/main/src/signals/numerical_adapter/src/shaders/double_slit.vert.glsl) and [double_slit.frag.glsl](https://github.com/westurner/sustainablefactory/blob/main/src/signals/numerical_adapter/src/shaders/double_slit.frag.glsl), parameterized by `num_slits > 0`, `slit_width`, `slit_separation`, `barycentric_shift`, `delta_oam`, and `azimuth_phi`. 5. **Revised Semi-Dirac Soliton Bus, Slow-Light CROW Delay Lines, Tiered 128-to-64-Bit RDMA LUT-CORDIC QRAM, & Integrated Multi-Band ($405\text{–}1550\text{ nm}$) CNC / $c\text{-BAs}$ Photonics (`semi_dirac_soliton_bus_crow_rdma_and_multiband_cnc_cbas_photonics_identities`)**: In [grassmannian-semi-dirac-soliton-bus-crow-rdma-qram-kicad-paper.md](grassmannian-semi-dirac-soliton-bus-crow-rdma-qram-kicad-paper.md) (**Paper 33**) and [grassmannian-multiband-1550nm-photonics-cnc-repeaters-cbas-homodyne-kicad-paper.md](grassmannian-multiband-1550nm-photonics-cnc-repeaters-cbas-homodyne-kicad-paper.md) (**Paper 34**), we refine the Soliton Bus with: - **$532\text{ nm}$ LCS Strain-Locked Semi-Dirac Ring-Core & $C_8$ Star-Via Channels**: Enforcing $v_{g,x}(0, m_x^*) = 0$, dispersionless $\hat{y}$ soliton propagation $\sigma_y^2(t) = \sigma_{y,0}^2$, $100\times$ RIE sidewall $D_{\mathrm{it}}$ suppression, and GHz ballistic Klein switching ($\mathcal{R}_{\mathrm{Klein}}(10) = 101$), - **$\mathrm{CO}_2$-Carbonate $5\times$ $\chi^{(3)}$ `CPHASE`/`CNOT` Gate Compression (`co2CarbonateChi3CphaseLength`)** & **Slow-Light `N-LIG` CROW Delay Lines (`crowSlowLightDelayLength`)**: Compressing the $24\text{ ns}$ RFSoC CORDIC buffer from $4.8\text{ m}$ ($n_g = 1.5$) down to $24\text{ mm}$ ($n_g = 300$), - **Tiered 128-to-64-Bit QNIC RDMA Header Stripping (`tieredQnicStrippedPayloadBits`), LUT-Seeded CORDIC Thermal-Drift Cancellation (`lutPlusCordicOamPhase`), & $37.5\%$ PWM 4K Cryo Budget (`nligInternalPumpAvgPowerMw`)** ($105.47\text{ mW} < 110\text{ mW}$ across 250 nodes), - **Periodic $10\text{ m}$ / $100\text{ m}$ Cellulose Nanocrystal (CNC) $\chi^{(2)}$ PSA Signal Boosters (`cncPeriodicBoosterNetLinkGainDb`, `cncCheckpointCount`)**, **$590\text{ nm}$ $n$-Adamantane NV Purcell Dowsing (`nvPurcellFactorNormalized` $> 12$)**, and **$1550\text{ nm}$ Inverted 3D Adiabatic Scarf-Joint Tapers into Cubic Boron Arsenide ($c\text{-BAs}$) Homodyne Detectors (`adiabaticScarfTaperLoss < fresnelButtJointReflectance`)**, with verified KiCAD v8 (`.kicad_mod`, `.kicad_sch`, `.kicad_pcb`, `.kicad_vol`), GDSII (`.gds`), and IOF/PROV-O (`_CBOM.yaml-ld`) hardware artifacts in `src/signals/kicad/semi_dirac/`. ### Scholarly References Travers, J. C., Grigorova, T. F., Brahms, C., and Belli, F. (2019), [High-energy pulse self-compression and ultraviolet generation through soliton dynamics in hollow capillary fibres](https://doi.org/10.1038/s41566-019-0416-4). Experimental hollow-capillary self-compression and UV dispersive-wave generation. Travers, J. C. (2024), [Optical solitons in hollow-core fibres](https://doi.org/10.1016/j.optcom.2023.130191). Review of gas-filled hollow-core propagation, self-compression, Raman shifting, photoionization, plasma, and dispersive-wave effects. Kivshar, Y. S., and Quiroga-Teixeiro, M. L. (1993), [Influence of cross-phase modulation on soliton switching in nonlinear optical fibers](https://doi.org/10.1364/ol.18.000980). Coupled nonlinear-fibre model for XPM-mediated soliton switching. Wang, A. et al. (2018), [Directly using 88-km conventional multi-mode fiber for 6-mode orbital angular momentum multiplexing transmission](https://doi.org/10.1364/oe.26.010038). The title says 88 km, while the abstract reports 8.8 km of OM4 fibre; the model follows the abstract and records 120-Gbit/s QPSK, six OAM mode groups, and 2x2 or 4x4 MIMO equalization. ## Evidence Boundary The implementation does not establish: - room-temperature soliton stability in N-LIG or rGO-Vitrimer; - preservation of a quantum state through a proposed material bus; - single-photon-strength Kerr coupling; - nondestructive parity measurement; - entropy siphoning or vacuum-noise removal; - a fault-tolerant QPU or QRAM architecture. QECLean was evaluated as a possible adapter. Its cached artifacts were built against a different mathlib context from Signals, so the public bus module uses a self-contained finite-matrix layer instead of making QECLean a hard dependency. The QECLean source remains available for future adapter work after toolchain alignment. ## Soliton-Bus Quantum Error-Coding Plan ### Recovered design choices The chat corpus search was case-insensitive literal `soliton` over `data/chats/*.md` only. `rg` was unavailable in the active terminal, so the equivalent `grep -i` search was used; it found 22 Markdown files. Detailed review focused on the processor/soliton and hardware-architecture chats cited below. These chats recover proposals, not experimental validation. These choices belong to separate axes: a design chooses how signals are addressed, how each code block encodes a state, and how checks are measured. They should not be collapsed into one meaning for “multiplexing.” 1. **Hierarchical optical addressing.** The proposal uses WDM as a coarse depth/layer selector, OAM or MDM as a finer channel selector, and a thermo-optic switch as an active route into an ancilla or memory layer. The chats also propose TDM to reuse a waveguide at different pulse times ([depth and OAM routing](https://github.com/westurner/sustainablefactory/blob/main/data/chats/_Quantum%20Processor%20and%20Soliton%20Discussions%20%20.md#L256), [time-bin reuse](https://github.com/westurner/sustainablefactory/blob/main/data/chats/_Quantum%20Computing%20Hardware%20Architectures%20Review%20%20.md#L1477)). **Use in v1:** represent these as address/scheduling coordinates only. Address separation does not prove low crosstalk or independent quantum channels. 2. **Port graph with an ancilla hub.** The proposed star has data leaves reached through a central ancilla, with bus operations intended to support stabilizer measurements ([star and bus-connected checks](https://github.com/westurner/sustainablefactory/blob/main/data/chats/_Quantum%20Processor%20and%20Soliton%20Discussions%20%20.md#L3172), [paper architecture](https://github.com/westurner/sustainablefactory/blob/main/docs/paper.md#L60)). **Use in v1:** call each physical endpoint a `BusPort`; represent possible interactions with an explicit coupling relation. A port is not itself a code qubit or a stabilizer. The measurement order, ancilla reuse, fault propagation, and photonic resource/measurement primitive still need specification. 3. **Multiplexed OAM qudit.** An alternative is to use a finite OAM basis as the data state, so mode shifts, phases, leakage, and mode mixing are data errors rather than just routing errors. The existing OAM-100 record is readiness metadata: the review explicitly says it lacks an OAM encoding map, syndrome readout, and recovery, and requires mode-dependent loss and crosstalk characterization ([OAM boundary and next measurements](https://github.com/westurner/sustainablefactory/blob/main/docs/paper-model-review.md#L1077), [QEC readiness fields](https://github.com/westurner/sustainablefactory/blob/main/docs/paper-model-review.md#L1137)). **Defer as a separate qudit-code project.** A coherent superposition over OAM modes is one qudit state, not a collection of independent addressed channels. 4. **Propagating GKP/CV state.** A grid state with homodyne syndrome measurements is a distinct continuous-variable encoding path. The paper review describes the cited optical work as a logical-state precursor, not completed fault-tolerant computation, and the qutrit/ququart break-even experiment is not an OAM experiment ([optical-state boundary](https://github.com/westurner/sustainablefactory/blob/main/docs/paper-model-review.md#L1071), [qudit experiment boundary](https://github.com/westurner/sustainablefactory/blob/main/docs/paper-model-review.md#L1098)). **Keep outside the first qubit/erasure API.** 5. **Layer-code geometry.** Layer codes are a CSS-to-3D topological-code construction: their layers and 1D junctions come from Tanner incidences. They are not optical wavelength layers or OAM channels. The paper’s construction supplies mathematical code properties, not a photonic decoder, threshold, or hardware benchmark ([Layer-code review](https://github.com/westurner/sustainablefactory/blob/main/docs/paper-model-review.md#L916)). **Treat this as a later code-family contribution**, independent of the optical address scheme. ### First implementation decision For the first implementation, WDM, MDM, OAM, polarization, and TDM are **transport/address dimensions**. QEC is defined over an abstract finite set of qubit blocks, independently of which optical degree of freedom eventually encodes each block. This matches the existing `LaneAddress`/`LaneSelector` boundary and lets QECLean prove qubit-code facts without making claims about N-LIG, OAM devices, or a quantum bus. OAM-as-data remains an explicit alternative encoding, never an implicit interpretation of `oamCharge`. Use **port** for the hardware endpoint and **address** for the multiplex selector: - `BusPort` identifies a coupler, interface, detector, or check-measurement endpoint. It is distinct from `LaneAddress`; multiple addresses may be routed through one port at different times. - `PortCoupling` (or a finite port graph) records which endpoints can interact and with what calibrated coupling/crosstalk. Do not define physical adjacency as consecutive `Fin` indices or infer it from OAM charge. - `CodeBlockPlacement` maps each abstract code-block ID to its required port and lane address(es). Require injectivity where resources must be distinct; model time-shared resources with an explicit schedule rather than pretending they are separate hardware. - An `EncodingChoice` states whether the eventual data encoding is an abstract qubit, a specified qubit encoding, an OAM qudit of dimension `d`, or a CV/GKP mode. The v1 Signals contract records the choice and its evidence status; it does not prove that a proposed encoding is physically realized. ### Signals work Add a focused `Signals/SolitonQEC.lean` module under `Signals.Pending`, reusing `Signals.SolitonBus.LaneAddress`, `BusNetwork`, and the existing unit and bounded-factor wrappers. Keep it out of the verified classical transport API until its fields describe calibrated observations or purely mathematical invariants. The deployment/noise contract should make these inputs explicit: - finite code-block and port indices, address-to-port placement, and the selected encoding interpretation; - a measured or assumed port-coupling graph, including which nearby signals can produce correlated faults; - mode-resolved loss with a distinction between **flagged erasure** and **unflagged loss**, plus detector efficiency and false/missed-flag rates; - mode crosstalk, phase/timing error, check-readout error, and correlated error support, each with units or normalization, uncertainty, calibration status, and provenance where available; - a syndrome-extraction/readout contract that says which check is measured, by what photonic primitive, with what outcome and disturbance bounds. Add focused examples for unique and colliding addresses, two addresses routed through one scheduled port, a flagged versus unflagged loss, and a correlated neighbor fault. These tests validate record laws and reject missing assumptions; they do not establish QEC performance. In particular, the existing XPM phase law is not a QND parity measurement, and the proposed room-temperature/no-dispersion, 10 GHz, and entropy-siphoning statements in `paper.md` remain design claims, not fields to promote to verified facts ([paper architecture and claims](https://github.com/westurner/sustainablefactory/blob/main/docs/paper.md#L60), [paper-model-review evidence boundary](https://github.com/westurner/sustainablefactory/blob/main/docs/paper-model-review.md#L1)). ### QECLean contribution Add a dependency-free, qubit-stabilizer erasure API in a new logical-framework module such as `QEC/Stabilizer/Framework/Core/Logical/ErasureCorrection.lean`, exported through the existing `Core.Logical` and `Core` umbrellas. Its public input is a known erased support `E : Finset (Fin n)`; define erasure correctability by the absence of a nontrivial logical Pauli supported entirely inside `E`, including the project’s phase-equivalence convention. Prove the two useful interfaces: 1. The stabilizer-code erasure criterion: the Pauli error set supported in `E` is jointly correctable exactly when no nontrivial logical operator is supported in `E`. 2. The distance corollary: for a code of distance `d`, every known erasure set with `E.card < d` is correctable. This is the erasure bound; do not replace it with the unknown-error bound `2t < d`. Use the existing Pauli support, logical-operator, and code-distance abstractions. Add a small checked example with the existing `[[5,1,3]]` code showing correction of a single known erased qubit, plus a boundary example where an erased set contains a logical support. This theorem is a code property, not an erasure decoder, channel model, or hardware threshold. Keep the work independent of `Signals` and `Physlib`; QECLean’s Lake configuration already uses mathlib and its own Lean tooling, and this contribution needs no new dependency. The existing [Layer scaffold](https://github.com/westurner/sustainablefactory/blob/main/src/QECLean/QEC/Stabilizer/Codes/Layer.lean#L274) can use the API later, but deriving its defect stabilizers is a separate task. ### Cross-project handoff and gates Do not import either Lean library into the other. For the first cross-check, maintain one tiny documented fixture containing only the finite block IDs, address/port placement, X/Z check data, erased-block set, and code parameters. Signals checks placement and declared calibration/noise fields; QECLean checks the code and erasure theorem. Keep physical observations and units on the Signals side, not in QECLean. The implementation sequence is: 1. Add and test the Signals Pending deployment/noise contract. 2. Add the generic QECLean erasure criterion and distance corollary; test the `[[5,1,3]]` one-erasure case. 3. Instantiate one finite fixture on both sides without a package dependency. 4. Only after detector and channel calibration, add a measured noise model and assess a concrete syndrome circuit/decoder. Keep OAM-qudit codes, CV/GKP codes, a full Layer-code lift, room-temperature operation, threshold, scalable throughput, and thermodynamic cooling claims outside v1. For docs-only changes, build the Sphinx docs and run `git diff --check`. When code is added, also run the Signals build and QECLean’s focused Lean build. ## Validation From `src/signals`: ```bash lake build Signals SignalsTests ``` The focused tests cover multiplex selector matching, demultiplexing, Boolean NAND identities, parity, routing-map involutions, operator inventory counts, and matrix-vector application. The build still reports the repository's pre-existing Physlib `if_pos` deprecation warning.