# 3D rendering, CENTRAL 1 and tensor-shell implementation trace Reading edition · 18 October 2026. Complete technical findings and formulas are retained. Artifact names and line locators replace local storage paths; the exact original text and locator map remain in the accompanying source record. Date: 18 October 2026. Read-only source inspection of the retained repository, commit `98c6509dbebdbd17aa5d7a4fdfde092ac2eef9db`. No browser or full application was launched; four bounded pure-function checks were executed locally. Source labels identify the retained files and line locations; exact repository locations are kept in the source register. ## Finding The source contains real three-dimensional geometry and rendering code, including a central reference explicitly labelled **1**, a Möbius surface with six centre-to-face reference rays, imported addressed triangle surfaces presented as **Mesh / tensor shell**, and separate volumetric wave/vortex and central-spiral models. It would be incorrect to describe this estate as containing no 3D rendering. The inspected renderer does not itself derive a native tensor shell from the BOX construction. It either calculates a declared reference surface or renders already-supplied source geometry. The exact BOX → native tensor-shell transition remains unlocated in this inspected source set. This is a specific linkage/retrieval boundary, not evidence that the author's construction does not exist. ## 1. Central 1 and the three-dimensional reference surface - `spatial-entry.js:1–2` attributes the display to supplied `BrainkManifoldVisualizer.tsx`; that original named component was not located in this checkout. This establishes the adapter's stated provenance, not independent recovery of the original component. - `spatial-entry.js:20` defines the actual 3D parametrisation: - x = (1.35 + (v/2) cos(u/2)) cos(u) - y = (1.35 + (v/2) cos(u/2)) sin(u) - z = (v/2) sin(u/2) - Lines 23–26 project one anchor at conventional coordinates (0,0,0), draw the literal label `1`, and draw six rays to (±1.8,0,0), (0,±1.8,0), (0,0,±1.8). These are the six face-contact directions of the surrounding reference volume. The code does not construct its faces or use them to derive the Möbius surface. - Lines 27–29 render 25 longitudinal tracks, each with 150 segments, plus 66 transverse lines, using depth sorting. These are sampled 3D coordinates projected onto a 2D canvas, not a static image. - Line 21 first rotates the substrate about its X axis by `substratePhase`, then applies independent observer/camera rotations `rx`, `ry` and perspective. Lines 38–39 make substrate inspection distinct from following supplied snapshots; lines 83–87 handle camera movement. - `README.md:3` explicitly says central 1 is an identity, not an instruction to add one to every numerical component. Lines 11–16 state that substrate rotation is a presentation adapter and does not invent an evolution law or prove the Observer Theorem. This must be preserved when discussing CENTRAL 1. ### Supplied service loops and scale `spatial-entry.js:8–14` admits at most 512 unique supplied node identities with finite positions, positive radii and a declared origin. The shared extent is max(distance(node, origin) + radius). Lines 30–33 use the common scale 1.8/extent, subtract the shared origin and draw each supplied loop as a two-turn Möbius boundary. Thus relative scale belongs to the containing environment and supplied coordinates; there is no synthetic service-node substitution. The six reference contacts are not service nodes (`README.md:13`). ## 2. The actual “Mesh. Tensor shell. Globe.” interface `repository-atlas.mjs:12` is the direct tensor-shell UI evidence. It labels the imported mesh layer “Mesh / tensor shell,” keeps `mesh`, `grid` and `book` layers separate, and says source coordinates stay fixed when the observer rotates the view. Implementation: - `kex-origin.mjs:2–3`: `(-1+)` is retained as a losslessly classified origin token, with no arithmetic normalisation. - `kex-origin.mjs:4–6`: admits supplied 3D vertices, triangular faces, resolvable `faceOwners`, unique node IDs/addresses, node `whole:1`, origin `(-1+)`, signed axes, positions, parents and optional online state. It returns a structural clone rather than transforming coordinates. Only `mesh`, `grid`, `book` are recognised projection layers. - `atlas-runtime.js:11`: camera-only rotation and perspective projection. - Lines 20–21: selects a supplied geometry layer and renders each vertex at vertex + node position. No BOX-to-sphere or BOX-to-tensor formula occurs here. - Lines 23–26: nodes with `online !== false` are visible. Global triangles are admitted when their declared face owner is online, then rendered without modifying vertices. This is source-face admission, not continuous morphing. - `kex-origin.mjs:7`: counts undirected edge incidence. `closedEdgeIncidence` means there is at least one face, no edge used once and no edge used more than twice. Its explicit claim is only edge-incidence observation. This predicate alone does not establish a sphere, an orientable manifold, an embedding without self-intersection, or the native recursive closure law. Existing tests: - `tests/kex-origin.test.mjs:4–6`: supplied geometry/identity preservation; rejection of invalid axes, duplicates, parent cycles and unresolved owners; open triangle versus closed tetrahedron incidence. - `tests/origin-runtime.test.mjs:4,20`: a tetrahedron with faces owned by A/B is closed with both admitted; disabling B makes it open; camera changes and export retain the source positions and projection layers. These are existing test definitions, not a claim that the entire suite was rerun in this task. - `origin-runtime.json:60–72,134–147`: the October 8 implementation record explicitly removed manufactured sphere/interpolation geometry, retained source-face admission, and recorded that the exact author mesh → tensor → globe rule had not been identified in its inspected material. This is a historical scoped record, not a universal absence claim. Its reported 78/78 checks at lines 161–168 were not rerun here. ## 3. BOX, planar shells and volumetric references are distinct stages The local BOX comparison is currently two-dimensional: `model.js:8–9` implements f(q)=sRq+t and its inverse. `app.js:13–14` nests three independently identified bodies through child and containing frames, retains right-face anchors and declares the outer rationale unquantified. `origin-v2-9.txt:46,70–72` describes that relation and explicitly limits the comparison to recoverability. `model.js:10–11,30` separately implements planar norm shells and checks 8r square/4r diamond counts. `app.js:18–19` renders a comparison embedding, with curvature c, rather than native 3D shell closure. The paper's lines 84–88 distinguish norm changes, rotations and the paraboloid F(x,y)=(x,y,c(x²+y²)); lines 3549–3555 locate the outstanding native 3D shell linkage. The volume models are genuine 3D but separately declared references: - `volume.js:5`: 957 fixed sample identities in a cylindrical envelope; position = origin + B·local; scalar amplitude = exp(−2(y²+z²)) cos(2π(x/λ−ft)). Lines 24–36 explicitly keep original positions separate from a display-height encoding. Lines 21–22,29–30 distinguish camera orbit from declared XY/XZ/YZ observer planes. - `vortex.js:5`: 3,465 material identities in an axisymmetric regularised vortex. ω(r)=Γ[1−exp(−r²/a²)]/(2πr²), with core limit Γ/(2πa²); local position = (x₀+Ut,r cos θ,r sin θ), θ=θ₀+ωt. The `wavelength` and `frequency` keys mean circulation and core radius here, not wave wavelength/frequency. It explicitly disclaims native tensor-shell verification. Lines 22–27 separate camera and observer planes; lines 29–33 draw streamlines/material markers and a trace. - `vortex.js:4–8` additionally contains `canonicalSpiral`: r=Ru, θ=direction·2π·turns·u+phase, x=pitch·turns·u, y=r cosθ, z=r sinθ, sampled at 513 u values. This function is absent from the newer `vortex.js`; do not silently merge these edition APIs. It is a declared Archimedean spiral with optional axial lift, not an empirical vortex law. ## 4. Origin notation must remain edition-specific CENTRAL **1** at the spatial renderer's conventional (0,0,0), the whole-projection token **(-1+)**, the historical volume `U=1`, and current workbench **`.1`** must not be silently identified. `singular-origin.js:2–5` requires `.1` as an exact opaque string and treats numeric projection coordinates as separate. Current volume/vortex use `projection([1,1,1])` (`volume.js:2,5`; `vortex.js:2,5`). Their historical original volume uses `origin_identity:'U=1'` and numeric [1,1,1] (`volume.js:2`). These implementation conventions document versions; they do not by themselves authorise mapping the user's CENTRAL 1 construction to current `.1`. ## 5. Fresh bounded checks performed Four source-inspected, pure-function checks passed (Node; no browser/network/full application): 1. Whole-projection import retained origin `(-1+)`, whole 1 and position [7,8,9] exactly and kept `normalization_applied:false` for signed-axis classification. 2. A supplied tetrahedron had four faces, zero boundary/nonmanifold edges and closed incidence. Admitting only A's two faces produced four boundary edges and open incidence while retaining the same vertex array. 3. The exact renderer parametrisation satisfied P(0,v)=P(2π,−v) with maximum residual 4.5076527157131×10⁻¹⁶, and P(0,v)=P(4π,v) with residual 9.0153054314262×10⁻¹⁶ for five v values. This checks the sampled Möbius seam/two-turn return, not native tensor closure. 4. The separate central spiral produced 513 stable IDs, strictly increasing radius, maximum radius residual 4.440892098500626×10⁻¹⁶ and exact sampled reversal of z under handedness reversal. No new browser-render observation, live publisher snapshot, native BOX→shell fixture, physical evolution law, full tensor-field operator, measurement comparison or independent physical validation was established by these checks. The next necessary evidence is the specific authored transition rule or saved source geometry/fixture connecting BOX and CENTRAL 1 to the intended tensor shell; that rule should be tested in its own native domain rather than replaced by the Möbius or analytic-vortex references. ## 6. Additional recovered manuscript and image evidence These source records belong to the accompanying research package, outside the retained repository. ### The line → square → cube / BOX bridge is in the retained conversation `retrieved-native-recovered-complete.md:2317–2647` retains the challenge about the number line filling a grid and faithfully representing square roots, followed by a line → surface → volume exposition. Specific anchors: - Lines 2317–2329: the challenge and singular-origin grid interpretation. - Lines 2443–2481: n, n², n³ and n^d as dimension-dependent relational extents. - Lines 2483–2515: explicit box-in-box link, n²×n=n³, and line → surface → volume. - Lines 2527–2619: cube-closure cardinalities and 64=8²=4³ under different observer-governed dimensional definitions. - Lines 2621–2647: substrate versus representation and the norm-based shape comparison. This is useful source evidence for the intended dependency, but is conversational exposition rather than the executable BOX→tensor renderer. Its speaker roles must not be flattened: it includes a user challenge and an answering explanation. Its introductory correction (lines 7–11) also explicitly reclassifies transported-integer algebra elsewhere in the same record as comparison construction. The document must not be used to silently restore withdrawn native-arithmetic attribution. Source link retained in the capture: https://docs.google.com/document/d/1b3wASx0Q08waPsYyfUjS2LdLUbXNrHJ-jNolkHPsN7s/edit ### Existing shell rendering is also directly evidenced in the monograph I inspected the authorized recovered `KEX_BRAINK_MONOGRAPH_DRAFT_XXXVII.docx`, its complete paragraph extraction `monograph-all-paragraphs.json`, and three of its actual embedded images. The first heading is Draft XII because the document retains multiple drafts; it must not be described as only Draft XII. - Extraction paragraph indices 746–748: “3D Wake-Shell Persistence Geometry” and its explanation. The DOCX embeds `image13.png` through relationship `rId21`, immediately after that heading (top-level XML body paragraphs 184–186). The visually inspected image is an actual 3D surface plot of a tubular/cavity shell with X/Y axes and a translation-axis coordinate. It is not merely a caption asserting a render exists. - Extraction indices 749–751: observer-state translation-distortion discussion. `image14.png`, relationship `rId22` (top-level body image paragraph 188), actually shows two one-dimensional curves labelled Source field and Observer-shifted field. Its role is a shifted readout comparison, not a second 3D shell plot. - Extraction indices 815–817: “Direct Representation Overlay — Translation and Residue.” `image19.png`, relationship `rId27` (top-level body image paragraph 254), visibly contains nested blue shell outlines, a red translation axis, trailing green arcs and a yellow Observer-state marker. The accompanying paragraph explicitly assigns those roles. - Extraction index 813 preserves open requirements: real-unit calibration, magnetohydrodynamics, radiative transfer, uncertainty propagation, distributed topology validation and statistical controls. It distinguishes recursive coherence from calibrated empirical validation. - Extraction index 988 gives an explicit **1D proxy** trace: source sine/envelope, shift, cumulative difference residue, normalisation, gradient dispersion and coherence weighting. It calls the calculation a proxy requiring real units/data/constraints. It does not provide the formula generating the pictured 3D shell or an authored BOX→tensor map. Only these three embedded source images were extracted for inspection, to `monograph-render-184-image13.png`, `monograph-render-187-image14.png` and `monograph-render-253-image19.png`. No new image was generated or edited. This strengthens the evidence that the author's materials already contain substantive rendered shell representations. None of the three inspected images labels CENTRAL 1 or supplies the missing native transition algorithm.