# AGLA FIGURAE A / III VECTOR SOURCE PACKAGE Current v0.3 technical qualification evidence: `FIGURA_A_DUPLEX_RELEASE_QUALIFICATION_REPORT_2026-08-21.md`. ## Additive Figura A duplex candidate The package now also contains a parallel 0.3.0-DUPLEX-CANDIDATE / HUMAN_REVIEW_REQUIRED / NON-RUNTIME lineage. It preserves the supplied Illustrator SVG and every v0.2 derived artifact byte-for-byte. The operative candidate treats ROTA A as one ROTA with an outer absolute-noun wheel and an inner concretum citation-form locus wheel typed `CONTRACTED_CITATION_FORM_LOCUS`. Existing `CONTRACTED` wheel, locus, DOM, and predication-carrier IDs remain stable; the address does not impose one morphological gender on the complete series. Its structural inventory is 9 ABS loci, 9 CON loci, 9 radial axes, 36 symmetrical pair-key indexes, and 72 unevaluated ABS_X-to-CON_Y predication carriers. The indexes are not Figura III camerae; the candidate materializes zero camerae and does not claim Stack G authority. Executed state belongs only to `carrier_result_records` (`CARRIER_RESULTS`), keyed by `[carrier_id, context_id, execution_id]`. OPERA lanes are non-primary overlays that reference existing carriers and, when bound, resolve to those records. The nine inner-object summaries are read-only `BY_PREDICATE` derivations, not independently writable locus state. `CARRIER` is reserved in the v0.3 vocabulary for those 72 directed `PREDICATION_CARRIER` edges. The nine inner objects are loci, never carriers. The witnessed inner series is BONUM MAGNUM DURANS POTENS SAPIENS VOLENS VIRTUOSUM VERUM GLORIOSUM. These are citation/display forms, not runtime surface outputs: the `-NS` forms are gender-syncretic rather than exclusively neuter, and surface realization is contextual, non-mechanical, runtime-bound, and owned by `SATOR_A`. DURANS is primary; Durabilis is a traced legacy lemma, not an equivalent. Witness attribution/date metadata remains unresolved, and human visual review remains required. The generated v0.3 SVG retains its labels as source-bound `LOCAL_PROMPT` `` elements under `../SVG_TEXT_LOCAL_PROMPT_LAW.md`; automated outlining is forbidden. Source Code Pro is embedded to stabilize metrics and prevent overflow, so editable text is not itself a release blocker. The inner `E_cycle_con` is derived from the ordered `locus_refs` B-C-D-E-F-G-H-I-K with cyclic wrap: nine render/traversal steps, no new edge objects, inference, execution, or results. The primary carrier-result key is `[carrier_id, context_id, execution_id]`. A primary record always has the complete key plus AREPO and OPERA references; null context/execution is legal only on the `EXAMPLE_UNBOUND` presentation overlay, which resolves to no result. Any non-empty derived CON-locus summary must list the source carrier composite keys; every source must target that locus and share its context and execution. The default SVG projection is BASE only, with Focus and OPERA hidden rather than simultaneously asserted. Its UTF-8 declaration and bytes agree, manifest paths name their bases, and validation executes the five fixed Draft 2020-12 schemas. The v0.3 schemas close their top-level records and the explicitly enumerated critical state, carrier-result, derived-summary, and overlay surfaces. Other nested topology, lexicon, and manifest metadata remains extension-tolerant; this package does not claim that every nested object surface is closed. See FIGURA_A_DUPLEX_V0_3_0_CONTRACT.md. Build and validate with: ~~~powershell & '.\TOOLS\BUILD_FIGURA_A_DUPLEX_SEMANTIC_SVG.ps1' & '.\TOOLS\VALIDATE_FIGURA_A_DUPLEX_V0_3.ps1' ~~~ The v0.2 description below remains as the preserved predecessor-line contract. Version: `0.2.0-FIGURAE-A-III-VECTOR-SOURCE` Status: `VALIDATED NUMERIC SOURCE CONTRACT / HUMAN_REVIEW_REQUIRED / NO VECTOR RELEASE / NON-RUNTIME` This package is the source-side contract for the hand-authored `Figura A` and `Figura III` vector masters. It treats a semantic SVG ROTA as a dynamic dataset whose declared color, visibility, focus, and wheel-pose fields can be projected from an admitted OPERA state. It binds stable object names to house colors, integral mutual convertibility, derived directional contraction, chamber color composition, and UI state without making color or geometry a substitute for textual identity. Version 0.2.0 assigns primitive colors **only to the A dignities**. The shared B-K selectors do not carry universal colors across regimens: ```text A = primitive dignity colors T = future relational descriptors for chromatic operations over A colors Q = future interrogative coordinates describing their circumstances S = subject/situated configuration in which the visual relation is expressed ``` This is a proposed visual type model only. AREPO remains responsible for admission and OPERA for selection/computation; Q does not execute a color operation and S is not an output engine. Therefore `#0000FF` is the human-approved visual color of `AB / Bonitas`, not a universal color of selector `B`, `TB`, `QB`, or `SB`. The package is separate from `01_SVG_OUTLINES` and the ATQS icon release. Those collections contain font-derived notation assets; this one describes semantic figural documents with groups, an arrowless integral relation field, derived state overlays, rotation records, focus projections, and palette tokens. ## Breachable representation distance Human-facing and machine-facing representations should be the same artifact whenever their requirements permit it. When they cannot be identical, they must remain two declared projections of one semantic source rather than two independent truths. ```text +-> HUMAN_RENDER SEMANTIC VECTOR + STATE -| +-> MACHINE_VIEW / DATA_RENDER HUMAN_RENDER -- stable object ID + manifest + declared transform --> SOURCE MACHINE_VIEW -- stable object ID + manifest + declared transform --> SOURCE ``` This package calls that requirement **breachable representation distance**. Every meaningful visible object and every machine record must meet all of the following conditions: - they resolve to the same stable ASCII object ID; - numeric changes have one stored source value and a declared render formula; - presentation-only effects are named, partitioned from semantic carriers, and retain an explicitly declared composite order; - a machine-only field that matters to interpretation has a human-readable cue, label, legend, inspection view, or documented diagnostic; - a human-visible semantic distinction has a machine-readable field and may not depend only on pixels, visual resemblance, or undocumented layer order; - exports record enough provenance to reconstruct the bridge and detect loss. The preferred vector therefore serves both audiences: named groups, relation kind, conditional direction markers, color fields, tone fields, wheel poses, focus, visibility, and UI alpha remain inspectable while the same geometry remains directly legible. Separate data bitmaps exist only because raster flattening destroys that object model. Their manifest and aligned planes bridge the destroyed structure explicitly. The full principle and its acceptance tests are specified in `REPRESENTATION_BRIDGE_PRINCIPLE.md`. ## Authority boundary - `TENET A` supplies the nine A-dignity identities and the declared display order `B C D E F G H I K`. - `ROTAS A` supplies the Figure A topology and the base/overlay distinction. - `TENET G`, `ROTAS G`, and the registered Stack G LIBER supply the 36 unordered Figure III chamber identities. - This package supplies a human-authored **visual encoding** of those identities. It does not create dignities, chambers, convertibility results, contraction grades, execution authority, or doctrinal facts. - Version 0.2.0 records the human-ratified visual rule that the complete, integral Figure A renders all 36 unordered dignity pairs as fully mutually convertible, single, arrowless relations. A formal executable OPERA binding and any more specific canonical source locator remain pending; this package does not manufacture either. - No primitive palette is assigned to T, Q, or S in this version. Color is a visual discriminator, not a machine identifier. Every meaningful object therefore retains an ASCII house, chamber, unordered-relation, or directed-edge ID. Reverse lookup from rendered color to house/chamber is prohibited. ## Integral base and derived projections The integral `Figura A` base contains one relation object for every unordered pair of the nine dignities: ```text C(9,2) = 36 unordered relations relation-BC = relation-CB convertibility(B,C) = full mutual convertibility ``` These 36 relation objects are arrowless. At full mutual convertibility there is no operational distinction between `X->Y` and `Y->X` to display; the absence of direction is therefore information, not missing geometry. Direction appears only in a **derived contracted overlay** whose admitted state distinguishes the two orientations of a pair. Such an overlay may contain two independently named and toned lanes, `edge-X-to-Y` and `edge-Y-to-X`. It must not rewrite the integral relation as though the complete base had always contained two arrows. A **focus projection** is a separate view. It may isolate one dignity and its incident relations, dim or hide other objects, and rotate the presentation toward the 12-o'clock focus target. In v0.2.0, B satisfies the declared canonical-slot equation and the current raster is visually consistent; actual glyph-anchor measurement remains pending. Focus alone does not create contraction or direction. The approved first example is: ```text focus_house = B focus_anchor = TOP_12_OCLOCK ``` The focus and rotation rules are specified in `ROTATION_FOCUS_AND_ALIGNMENT_CONTRACT.md`. ## Three independent visual channels 1. **Hue / base RGB** identifies a house visually. 2. **Tone** represents a semantic degree: - an integral arrowless relation remains at full tone; - a derived directed edge uses its directed convertibility; - a dignity uses its aggregate net convertibility; - contraction or loss is the complement of that value. 3. **Alpha** is reserved for UI selection and on/off state. It must not encode contraction or convertibility. Canonical scalar storage uses integer parts per million (`ppm`) from `0` to `1,000,000`: ```text integral_relation_tone({X,Y}) = 1,000,000 derived_edge_tone(X->Y) = directed_convertibility(X->Y) derived_edge_loss(X->Y) = 1,000,000 - directed_convertibility(X->Y) outgoing(X) = mean(convertibility(X->Y), Y != X) incoming(X) = mean(convertibility(Y->X), Y != X) net_convertibility(X) = mean(outgoing(X), incoming(X)) contraction(X) = 1,000,000 - net_convertibility(X) node_tone(X) = net_convertibility(X) ``` In version 0.2.0 these percentages form a `HUMAN_DIRECTED_VISUAL_METRIC`, not an output of the currently registered OPERA A. Its values are synthetic examples and its equal-weight aggregation is provisional. The term `contraction` here is a visualization complement; it does not define or grade doctrinal contraction. The existing example comparison matrix contains eight outgoing and eight incoming values per dignity and is retained only as a synthetic **derived contracted-overlay fixture**. Its older direct/analytic classifications do not define the 36-object integral visual base. Before non-synthetic use, every directional value still requires the governing mediation and OPERA evidence. Missing evidence never silently means zero; a net visual score is unavailable until coverage satisfies its declared aggregation policy. ## Dynamic ROTA dataset Composing the SVG is itself the construction of a dynamic ROTA dataset. The geometry is stable; an external controller applies a versioned state record to named objects. An applied state may change only fields it explicitly governs, including: - base/overlay visibility and UI emphasis; - dignity and relation `tone_ppm` values; - focused dignity and focus projection; - global view alignment; - declared semantic rotations between wheels; - presentation blending after semantic fields are materialized. The SVG must not infer an OPERA from its current appearance. OPERA admits or computes the state; the SVG expresses that state and retains its source reference. Until the relevant OPERA locator and schema are registered, use a nullable `opera_state_ref` plus an explicit `PENDING_BINDING` status rather than inventing authority. Whole-view rotation used to place the focus at 12 o'clock is presentational and does not alter relative wheel positions. Rotation of one wheel relative to another is a distinct semantic pose and may express an OPERA stage only when the state record says so. These transforms must never share one unnamed SVG `transform` value. ## Tone rendering For the color rendering profile, tone scales the encoded sRGB channels with one final half-up rounding step: ```text render_channel_8 = floor((base_channel_8 * tone_ppm + 500,000) / 1,000,000) ``` This is equivalent to reducing HSV/HSB Value while retaining the base hue and saturation ratios. The word `luminosity` in authoring conversation maps to the model-independent semantic field `tone_ppm`; it must not be confused with HSL Lightness, CIE L*, relative luminance, or opacity. For a palette-independent grayscale export: ```text gray_8 = floor((255 * tone_ppm + 500,000) / 1,000,000) ``` This neutral grayscale profile is the comparison-safe fallback. Converting the colored artwork to grayscale ad hoc is not comparison-safe because different hues have different intrinsic luminance. ## Figure III chamber color The current human-supplied `BC`/pair design is recorded with a provisional baseline profile using saturated addition of the two opaque base sRGB8 colors: ```text mix_channel_8 = min(255, left_channel_8 + right_channel_8) ``` Algorithm ID: `AGLA.SRGB8_SATURATED_CHANNEL_ADD.V1`. This is a deterministic design operation over encoded sRGB8 channels. It is not alpha compositing, pigment mixing, perceptual interpolation, or a claim about physical light. If a chamber is later toned, composition order is `MIX_BASE_COLORS_THEN_APPLY_TONE`. This baseline is not a universal chamber-combination law. A future T-mediated color-combinatorics layer may select or parameterize interaction operators from declared T relations. The design note is recorded in `FUTURE_T_COLOR_COMBINATORICS.md`; no such mapping executes in version 0.2.0. Because alpha is reserved, an overlapping chamber must use three opaque regions: the first component-only region, the second component-only region, and an explicit intersection region filled with the materialized additive color. Do not obtain the intersection color by making two circles translucent. ## Supplied SVG candidate `ROTAE STRUCTURE AND ROTA A.svg` is now present as a human-authored source candidate and has been structurally inspected: ```text artifact status : SUPPLIED / STRUCTURALLY_INSPECTED_CANDIDATE SHA-256 : 4D17F4704AF3793B7D6246E324F876763C4919BD6F9C15CF087912DE8457D972 viewBox : 0 0 1920 1920 generator : Adobe Illustrator 30.7.0 ``` Observed groups include `ROTA_A`, `ROTA_A_WHEELS`, `ROTA_A_HOUSES`, and `ROTA_A_CONNECTIONS_x5F_cuspiid`. All nine per-focus connection groups are present. The `B_x5F_CONNECTIONS` group is visible in the supplied snapshot; the C-K groups use the current `st5` hidden class. This is strong structural evidence for the approved B-focus view, but it is not an OPERA binding. Illustrator escaped underscores in several IDs as `_x5F_`. Those IDs and the current display/opacity classes require a declared manifest mapping before release. The focus connection groups also contain arrowed appearance geometry. In an integral focus snapshot those arrows are traversal/navigation cues, not asymmetric convertibility; they become semantic direction only in a separately bound contracted overlay. They are not proof that the integral base consists of directed relations. The byte-preserved source candidate still lacks the explicit 36-relation taxonomy and state bindings; these are materialized only in the generated semantic derivative described below. The candidate's native 1920-square coordinate system must be preserved. A normalized viewBox, if later required by an application, is a declared derived adapter and must not silently rewrite the supplied source geometry. ## Generated semantic candidate The deterministic semantic build now exists with truthful status `GENERATED_VALIDATION_PENDING / NON_RUNTIME`: ```text semantic SVG path DERIVED/semantic/AGLA_FIGURA_A_ROTA_SEMANTIC_v0_2_0.svg bytes 372501 SHA-256 BBB676DCC592AB85DE8A6D0C61D93A4D4B9AD49767FDB4AECB9873CC42A43B1A manifest path DERIVED/manifests/AGLA_FIGURA_A_ROTA_SEMANTIC_v0_2_0.manifest.json bytes 14598 SHA-256 407EFD938B276532D07CDDCF7DB270D4C4CCEB6A3A76CCA93F544823274EBC27 B-focus semantic preview path DERIVED/preview/ROTA_A_SEMANTIC_v0_2_0_FOCUS_B.png bytes 257899 SHA-256 F5FA7270F5CE9F35B7DDEB8CE3A00FDC03E94037EC1FC23CADB5D140EBC6DE22 source-render comparison path DERIVED/preview/ROTA_A_SOURCE_RENDER.png bytes 281662 SHA-256 362CAA0671900B1E6B19CC74B97F28E7E7C75098AD5FF1A83C88266722F47B8C ``` The two PNG comparison artifacts intentionally differ. In the semantic candidate, the palette identity band emits the exact thirds colors directly, at semantic opacity 1, without the Illustrator blend-mode opacity that affected the source composite. The source white house wedges and wheel/center appearance retain their original opacity only in the co-rotating, nonsemantic `wheel-presentation-underlay`; the unbound tone preview is hidden. B focus and its eight incident UI lanes remain present, while contraction tone is materialized only in the hidden, state-bound node and directed-edge overlays. This is a declared source-to-semantic visual difference, not renderer drift. It does not by itself validate textual outlines, OPERA binding, raster data-plane semantics, or runtime authority. The generated wheel subtree is: ```text semantic-wheel-transforms wheel-presentation-underlay presentation/nonsemantic; source opacity retained canonical-base semantic carriers; opacity 1 houses integral-relations semantic-state-projections contracted-node-tone-layer contracted-state-overlay wheel-presentation-overlays presentation/nonsemantic ui-focus-projection ``` The underlay, canonical base, state projections, and presentation overlays are siblings under `semantic-wheel-transforms`; only `ui-focus-projection` is a child of `wheel-presentation-overlays`. This keeps all co-rotating appearance geometry in the correct transform space without allowing it to become machine-authoritative semantic color or opacity. The manifest reports: ```text houses 9 integral relations 36 integral arrowheads 0 bound focus states 1 unbound source lane sets 8 focus traversal lanes 72 contracted dignity nodes 9 contracted example lanes 72 SVG elements 1292 SVG IDs 397 unique SVG IDs 397 ``` Default reproduction from the package root is: ```powershell & ".\TOOLS\BUILD_FIGURA_A_SEMANTIC_SVG.ps1" ``` The build reads the byte-preserved source SVG, the integral-base topology, and the example state; writes the exact semantic/manifest paths above; preserves the `0 0 1920 1920` viewBox; and emits no generation timestamp. The manifest records `runtime_authority=false`, with AREPO and OPERA bindings still pending. The manifest visual-audit status remains `HUMAN_REVIEW_REQUIRED`. Human audit remains mandatory for the outlined dignity text. Raster inspection cannot safely settle the `MAGNITAS`/canonical `MAGNITUDO` question, confirm the outlined `POTESTAS` form without editable source text, or resolve the D-label ambiguity. No outlined geometry was silently rewritten. ## Source files - `ROTAE STRUCTURE AND ROTA A.svg` — supplied, structurally inspected SVG candidate; not yet the validated semantic interchange release. - `SOURCE/tokens/agla_figurae_a_iii.palette.tokens.json` — nine house colors, half-tone references, all 36 chamber colors, and the render contracts. - `SOURCE/states/agla_figura_a.state.example.json` — synthetic 72-direction derived-overlay fixture showing asymmetric `B->K` and `K->B` convertibility; it is not the integral base. - `SOURCE/topology/agla_figura_a.integral-base.json` — explicit nine-house, 36-unordered-relation integral topology used by the semantic builder. - `SOURCE/projections/agla_figura_a.focus-b.projection.example.json` — B-focus UI traversal/alignment projection, independent from convertibility direction. - `SCHEMAS/` — five Draft 2020-12 JSON Schemas for palette, integral base, contracted state, focus projection, and generated semantic manifest records; closure applies at their declared boundaries, excluding intentional payload/value extension surfaces. - `DERIVED/semantic/AGLA_FIGURA_A_ROTA_SEMANTIC_v0_2_0.svg` — generated semantic SVG, validation pending and non-runtime. - `DERIVED/manifests/AGLA_FIGURA_A_ROTA_SEMANTIC_v0_2_0.manifest.json` — generated provenance, taxonomy, counts, hashes, and audit status. - `DERIVED/preview/` — source/semantic comparison rasters and the B-focus preview; presentation evidence, not `DATA_RENDER` planes. - `VECTOR_AUTHORING_CONTRACT.md` — Illustrator/SVG layer, group, path, and ID rules. - `RASTER_EXPORT_PROFILES.md` — separate human-readable and machine-readable bitmap contracts derived from the same semantic vector. - `FUTURE_T_COLOR_COMBINATORICS.md` — non-executing research register for relation-sensitive color composition. - `DYNAMIC_ICON_COLOR_BINDING.md` — monochrome reusable icon geometry plus explicit palette, tone, and UI-state bindings. - `ROTATION_FOCUS_AND_ALIGNMENT_CONTRACT.md` — 12-o'clock focus invariant, integral/derived view modes, and separation of view alignment from semantic interwheel rotation. - `PORTABLE_SCENE_ADAPTER_CONTRACT.md` — engine-neutral scene binding for future Godot, Unreal, and other 2D/3D adapters. - `TOOLS/BUILD_FIGURA_A_SEMANTIC_SVG.ps1` — deterministic source-to-semantic SVG and manifest generator. - `TOOLS/SYNC_FIGURAE_A_III_DERIVED_VALUES.ps1` — deterministic recomputation of approved palette thirds, tone colors, chamber colors, collisions, and example state projections. - `TOOLS/VALIDATE_FIGURAE_A_III_VECTOR_PACKAGE.ps1` — PowerShell 5.1 semantic validator. - `TESTS_FIGURAE_A_III_VECTOR_PACKAGE.md` — test and acceptance registry. `SOURCE/illustrator/` is reserved for the human-authored `.ai` master. The Illustrator file is an editable geometry source; the JSON tokens and state records remain the exact numeric record. `DERIVED/` is reserved for exports and must be reproducible from the source master plus these contracts. The semantic vector may be the common source for both presentation and data use because its semantic color, tone, UI alpha, stable IDs, and companion records remain inspectable. Semantic carriers are opaque. Retained source opacity and blending belong to the explicitly nonsemantic presentation underlay and may change the displayed composite without changing stored fields. A bitmap has no such retained object separation after flattening. It must therefore be exported through either the explicitly aesthetic `HUMAN_RENDER` profile or the lossless-field `DATA_RENDER` profile; the two bitmaps are not interchangeable. ## Current palette construction The user clarified that Illustrator interpolated the shown sequence from pure Blue, Red, and Green and authorized an exact reconstruction rather than screenshot sampling. Each primary-to-primary segment is divided into three equal componentwise intervals in encoded sRGB8: ```text B #0000FF -> C #5500AA -> D #AA0055 -> E #FF0000 E #FF0000 -> F #AA5500 -> G #55AA00 -> H #00FF00 H #00FF00 -> I #00AA55 -> K #0055AA -> B #0000FF ``` The exact channel step is `255 / 3 = 85`; base colors require no rounding. Half-tone references, provisional chamber mixtures, and example state colors are materialized by the synchronizer and checked by the validator. The numeric construction is human-approved while remaining a non-runtime visual encoding. The editable `.ai` remains `NOT_SUPPLIED`. The semantic SVG, provenance manifest, source comparison raster, and B-focus preview are `GENERATED_VALIDATION_PENDING / NON_RUNTIME`. They are not a validated release. EPS/ASE derivatives, formal `HUMAN_RENDER` release assets, `DATA_RENDER` plane bundles, portable scene adapters, and a runtime-authoritative release remain `NOT_GENERATED` pending human audit, semantic validation, and authoritative state binding.