{
  "title": "Otterchip Model E \u2014 independent final-geometry review (frozen compact E)",
  "date": "2026-11-06",
  "reviewer": "Claude Opus 5.5 (claude-opus-5-5), effort high; independent reviewer writing only hardware/otterchip_model_e/review/**",
  "verdict": {
    "geometryAndVisual": "No blocking error found. Within the app's estimated-body solver, the frozen compact E has zero hard flags in the cases I re-checked: reach, arm contact, knee, feet, the white-key-target bridge check, and black-key exposure. A secondary black-contact check is also reported. Raw forearm, wrist and shoulder metrics improve over M at every age, and the 70/30 aggregate and typing loss improve at every age. Music-only loss is worse than M at nominal ages 3\u20137. Two tradeoffs remain. At age 3, the screen top sits slightly above the reference eye and gaze is shallow. At age 12, gaze is steeper than the adult reference.",
    "productionReadiness": "NOT production-validated. Build 9 export geometry checks pass (see review-manufacturing.json, QA 2026-11-06T05:24:51Z). Stage A coupon and fit files can be prepared for quotation and printing. The full shell is conditional on measured fit, coupons and the acoustic prototype. No physical sit, load, tip, thermal, electrical, acoustic or child-use testing and no FEA have been performed.",
    "notAClinicalClaim": "J and every component score are designer heuristics computed on age-interpolated, historically estimated body proportions with CDC-midpoint stature. A lower J, or the M\u2192E J ratio, must not be presented as a percentage improvement in health, comfort or fitness."
  },
  "finalParameterIdentity": {
    "frozenByRoot": {
      "width": 355,
      "rear": 92,
      "height": 245,
      "tilt": 18,
      "setback": 180,
      "deck": 62.5,
      "slope": 0,
      "mididrop": 23,
      "midioverlap": 39,
      "padrise": 7,
      "padcrown": 1,
      "padgap": 10,
      "round": 4,
      "grillecorner": 4
    },
    "evidence": [
      {
        "file": "review/native-browser-final.json",
        "captured": "2026-11-06T04:53:43Z",
        "errors": [],
        "matchesFrozen": true,
        "note": "All 16 cases (M/E \u00d7 ages 3,6,9,12 \u00d7 typing/music) carry the frozen E params, including rear 92. E overall bounding box is 355 \u00d7 338.0 \u00d7 415.4 mm (W \u00d7 H \u00d7 D). M is 355 \u00d7 316.9 \u00d7 382.2 mm."
      },
      {
        "file": "optimization/native/selected-check.json",
        "matchesFrozen": true,
        "note": "main.params shows rear 92, height 245, tilt 18, setback 180. Aggregate J now 0.014797756266410417 (final core 1.1.0, nominal profile path; an earlier selected-check snapshot showed 0.0147936 from a tiny stature-rounding path difference), worst age 3, hardPenalty 0. Earlier artifacts (grille-alignment.json, the provisional captures, and the depth-grid rows) used rear 100 and/or height 250 / tilt 20 / setback 159. They are labelled provisional evidence below."
      }
    ],
    "status": "Frozen practical compact tradeoff. The optimizer's broader screen grid found a lower J at setback 250, so this is NOT a global optimum."
  },
  "imagesRead": [
    "/tmp/otterchip-e-final-E-body.png",
    "/tmp/otterchip-e-final-M-body.png",
    "/tmp/otterchip-e-review-E-orthographic-front.png (provisional geometry; superseded by the final orthographic image)",
    "/tmp/otterchip-e-final-E-3-typing-side.png",
    "/tmp/otterchip-e-final-E-3-music-side.png",
    "/tmp/otterchip-e-final-E-6-typing-side.png",
    "/tmp/otterchip-e-final-E-6-music-side.png",
    "/tmp/otterchip-e-final-E-9-typing-side.png",
    "/tmp/otterchip-e-final-E-9-music-side.png",
    "/tmp/otterchip-e-final-E-12-typing-side.png",
    "/tmp/otterchip-e-final-E-12-music-side.png",
    "/tmp/otterchip-e-final-E-6-typing-threequarter.png",
    "/tmp/otterchip-e-final-E-6-music-threequarter.png",
    "/tmp/otterchip-e-final-M-3-typing-side.png",
    "/tmp/otterchip-e-final-M-3-music-side.png",
    "/tmp/otterchip-e-final-M-12-typing-side.png",
    "/tmp/otterchip-e-final-M-12-music-side.png",
    "/tmp/otterchip-e-final-E-orthographic-front.png (frozen 245/18/180/rear 92; grilles visibly share the display top and bottom lines)"
  ],
  "rawMetrics": {
    "source": "native-browser-final.json (app solver, auto furniture with the same algorithm for both models). View distance, gaze, incidence and screen-top values use device.screenCenter, device.screenNormal and solution.points.eyes; they match native-final-equivalence.json raw values (e.g. E age 3 typing incidence 7.66\u00b0, view 386.1 mm). Horizontal shoulder\u2192wrist is my recomputation from exported points. Table heights match M within 2 mm at every age.",
    "definitions": {
      "horizShoulderWristMm": "Horizontal (XZ) distance from the solved right shoulder joint to the right wrist centre. This is the raw quantity the root graph plots.",
      "wristDisplacementRatio": "Optimizer loss input: |W \u2212 W0| / (upper arm + forearm), where W0 is the neutral wrist (upper arm vertical, forearm level). The good threshold is 0.243 (25\u00b0 upper-arm flexion equivalent).",
      "descriptiveChainRatio": "My earlier straight-line shoulder\u2192wrist \u00f7 (upper arm + forearm + hand) is a separate descriptive metric only. I do not ask for any graph change.",
      "tableLiftOverNeutralHandMm": "Solver field keyAboveNeutralMm = auto table height \u2212 neutral-hand table height for this seat. The neutral-hand construction is: upper arm vertical, forearm level, wrist 0\u00b0, palm level, and the activity's curled-finger fingertip drop down to the key target. The value is how far the key target (and therefore the table) sits ABOVE that construction. The auto policy accepts this lift to keep 20 mm thigh/knee clearance under the table. It is NOT the key height above the elbow joint. Read the raw forearm slope and wrist angle for the arm consequence."
    },
    "columns": [
      "model",
      "age",
      "activity",
      "viewMm",
      "gazeDeg",
      "neckDeg",
      "incidenceDeg",
      "screenTopBelowEyeMm",
      "elbowDeg",
      "wristDeg(+ext)",
      "forearmDeg(+rising)",
      "shoulderFlexDeg",
      "tableLiftOverNeutralHandMm",
      "horizShoulderWristMm",
      "wristDisplacementRatio",
      "penetrationMm"
    ],
    "rows": [
      [
        "M",
        3,
        "typing",
        379.6,
        14.2,
        0,
        3.7,
        12.7,
        93.9,
        -8.5,
        27.7,
        31.6,
        112,
        202,
        0.377,
        0.3
      ],
      [
        "E",
        3,
        "typing",
        386.1,
        10.3,
        0,
        7.7,
        -8.7,
        95.5,
        -2,
        19.5,
        25,
        83,
        193,
        0.291,
        0
      ],
      [
        "M",
        3,
        "music",
        437,
        12.9,
        0,
        2.4,
        16.7,
        89.2,
        1,
        11.9,
        11,
        54,
        159,
        0.141,
        0
      ],
      [
        "E",
        3,
        "music",
        438.8,
        9.5,
        0,
        8.5,
        -5.7,
        88.9,
        1,
        6.6,
        5.5,
        30,
        145,
        0.073,
        0
      ],
      [
        "M",
        6,
        "typing",
        460.7,
        20.3,
        0.3,
        9.8,
        78.9,
        98.3,
        -3,
        20.3,
        28.6,
        111,
        250,
        0.327,
        0
      ],
      [
        "E",
        6,
        "typing",
        441.4,
        17.7,
        0,
        0.3,
        55.9,
        90.6,
        1,
        15.1,
        15.7,
        84,
        211,
        0.192,
        0
      ],
      [
        "M",
        6,
        "music",
        500.7,
        18.9,
        0,
        8.4,
        81.3,
        88.9,
        1,
        9.7,
        8.5,
        53,
        189,
        0.111,
        0
      ],
      [
        "E",
        6,
        "music",
        499.8,
        15.9,
        0,
        2.1,
        58.9,
        88.8,
        0.5,
        5.5,
        4.3,
        29,
        176,
        0.059,
        0
      ],
      [
        "M",
        9,
        "typing",
        486.3,
        24.8,
        4.8,
        14.3,
        123.6,
        89,
        1,
        18.9,
        17.9,
        120,
        254,
        0.226,
        0
      ],
      [
        "E",
        9,
        "typing",
        491.3,
        22.1,
        2.1,
        4.1,
        106.8,
        89,
        1,
        14.7,
        13.7,
        91,
        241,
        0.174,
        0
      ],
      [
        "M",
        9,
        "music",
        555.5,
        22.5,
        2.5,
        12,
        131.6,
        89.2,
        1,
        9.2,
        8.3,
        59,
        223,
        0.107,
        0
      ],
      [
        "E",
        9,
        "music",
        560.3,
        19.9,
        0,
        1.9,
        113,
        89.2,
        0.5,
        5.6,
        4.8,
        35,
        209,
        0.063,
        0
      ],
      [
        "M",
        12,
        "typing",
        533.2,
        28.8,
        8.8,
        18.3,
        176.4,
        89,
        1,
        16.7,
        15.7,
        118,
        282,
        0.198,
        0
      ],
      [
        "E",
        12,
        "typing",
        536.3,
        26.2,
        6.2,
        8.2,
        159,
        89.1,
        1,
        12.9,
        12,
        90,
        268,
        0.153,
        0
      ],
      [
        "M",
        12,
        "music",
        601,
        26.2,
        6.2,
        15.7,
        184.4,
        89,
        0.5,
        8.1,
        7.2,
        57,
        249,
        0.093,
        0
      ],
      [
        "E",
        12,
        "music",
        602.3,
        23.9,
        3.9,
        5.9,
        166.1,
        89,
        0.5,
        4.6,
        3.6,
        32,
        233,
        0.05,
        0
      ]
    ],
    "age3vs12_E": {
      "typing": {
        "age3": "wrist \u22122.0\u00b0, elbow 95.5\u00b0, view 386 mm, gaze 10.3\u00b0, neck 0\u00b0, incidence 7.7\u00b0, screen top 8.7 mm ABOVE eye, shoulder\u2192wrist 193 mm horizontal, forearm +19.5\u00b0",
        "age12": "wrist +1.0\u00b0, elbow 89.1\u00b0, view 536 mm, gaze 26.2\u00b0, neck 6.2\u00b0, incidence 8.2\u00b0, screen top 159 mm below eye, shoulder\u2192wrist 268 mm, forearm +12.9\u00b0"
      },
      "music": {
        "age3": "wrist +1.0\u00b0, elbow 88.9\u00b0, view 439 mm, gaze 9.5\u00b0, incidence 8.5\u00b0, screen top 5.7 mm ABOVE eye, shoulder\u2192wrist 145 mm, forearm +6.6\u00b0",
        "age12": "wrist +0.5\u00b0, elbow 89.0\u00b0, view 602 mm, gaze 23.9\u00b0, neck 3.9\u00b0, incidence 5.9\u00b0, shoulder\u2192wrist 233 mm, forearm +4.6\u00b0"
      }
    }
  },
  "independentChecks": [
    {
      "id": "C1",
      "result": "verified",
      "title": "70/30 and composite arithmetic (canonical values from the native cross-check)",
      "detail": "Read core OBJECTIVE and composite():\n- activity weights 0.7/0.3, renormalised over the activities present, so D/W get typing-only scores with no imputed music;\n- category weights 0.25 each for view, elbow, wrist and reach;\n- view subweights: gaze 0.30, head tilt 0.20, incidence 0.25, distance 0.25;\n- J = 0.5\u00b7mean(nominal) + 0.25\u00b7mean(\u00b18% stature) + 0.25\u00b7max-age(nominal) + hard penalties.\nCanonical values come from native-final-equivalence.json (113 checks, 0 failures; browser geometry equals the optimizer adapter). Direct core.score on the captured E age-3 solution gives typing 0.02582332267241637 and music 0.024163303025379834. 0.7\u00b7typing + 0.3\u00b7music = 0.025325316778305407, exactly the selected-check age3 value. My own earlier hand recomputation (typing \u22480.0261, music \u22480.0240) came from separately calling component functions. It is approximate only and does NOT indicate any model discrepancy. The 4-age 70/30 subset means M \u22480.047 and E \u22480.014 are also approximate and only indicate direction; they are not J."
    },
    {
      "id": "C2",
      "result": "verified finding",
      "title": "What actually drives J",
      "detail": "In all 16 cases the wrist category is 0 and the elbow category is \u2264 0.006, because the solver's arm IK almost always lands inside those bands. Half of the category weight therefore barely separates designs. J differences come mainly from (a) wrist displacement at age 3 typing (M 0.30 \u2192 E 0.038) and (b) viewing at the age extremes (M age 12 typing view 0.41 \u2192 E 0.12; but age 3 typing 0.002 \u2192 0.066 and age 3 music 0.013 \u2192 0.091, both worse in E). Forearm pitch and table lift over the neutral-hand construction are not directly in J. Report them raw, as here."
    },
    {
      "id": "C3",
      "result": "approximate (earlier 4-age analysis); corrected by the final optimizer data",
      "title": "Activity and group weight sensitivity: superseded by the final optimizer per-age results",
      "detail": "My earlier 4-age recomputation was approximate and only indicated direction. The final optimizer results (finalOptimizerResults) show E lower than M for the 70/30 aggregate at every age, and for typing at every age. Music-only, however, E is WORSE than M at nominal ages 3\u20137 and better at 8\u201312. My earlier statement that E ranks ahead under every activity split is withdrawn for the music-only subgroup at ages 3\u20137."
    },
    {
      "id": "C4",
      "result": "verified",
      "title": "Hard and physical gates: white-key target and secondary black-key contact, reported separately",
      "detail": "WHITE-KEY TARGET: core.bridgeCheck on the final E music cases gives a minimum clearance of 40.0 mm (leftFingertip, ahead of the bridge face). The native 60-case result agrees. This is the representative white-key target only; it does not prove the whole playing area.\nSECONDARY BLACK-KEY CONTACT, using the mechanical corrected datum: the viewer reports 26.1 mm black-key exposure. In CAD, the flat black-key top begins at local z 15.6, the keybed centre is at 97.8, and the solid bridge/front lip is at 88.35, giving 25.05 mm actual CAD exposure. The earlier 28.55 mm figure was the keycap ramp envelope, not the flat playing top. The CAD lip extends 1.05 mm forward of the viewer bridge, so the optimizer's black-target capsule clearance of 9.7 mm becomes conservatively 8.65 mm. That is still above the 3 mm design allowance.\nOTHER: 25.05 mm stays above the 23 mm physical gate. The bridge underside sits 6.4 mm above the black-key tops (gate 3.4). Penetration is 0 in all 8 E cases. core.physicalChecks for E: none.\nThese are geometric checks of a static hand. They are not a claim that every key fits every hand. Donor measurement and real finger/playing trials are pending.\nNote: the same E-oriented physical check flags M's bridge gap at 3.36 mm against 3.4 mm (a 0.04 mm margin). That is informational only; M is unchanged and remains the default."
    },
    {
      "id": "C5",
      "result": "verified (one-variable only)",
      "title": "Neighbourhood robustness",
      "detail": "selected-check.json: 26 one-variable jitters (height, setback, deck, padgap, padrise, midioverlap, tilt at \u00b11/\u00b12) all have zero hard flags, with J between 0.01387 and 0.01866. Several neighbours score lower J: padrise \u22122 \u2192 0.01387, padgap +2 \u2192 0.01450, midioverlap \u22122 \u2192 0.01441, and the pad candidate gap 14.5/rise 6 \u2192 0.01392. The frozen point is therefore a practical rounded choice, not a local J minimum. That is acceptable, because the differences are small next to M's 0.05234. Multi-variable interaction and the hip-grid discontinuity were not exhaustively searched."
    },
    {
      "id": "C6",
      "result": "quantified",
      "title": "Display-depth choice vs numerical best",
      "detail": "From native-review-notes.txt and data/search.log:\n- screen-only grid best: 222/18/250, J 0.01070;\n- broader stage-2 best: setback 250, J 0.00920 (other params differ);\n- best with setback \u2264180: 244/18/180, J 0.01486; frozen rounded 245/18/180, J 0.01479 (screen-grid snapshot, 2026-11-05/06; the final full-solver J is 0.014797756266410417);\n- best with setback \u2264160: J 0.01655; M: J 0.05234.\nDepth: depth-grid rows give depth \u2248 setback + 246 mm at rear 100. Final E with rear 92 measures 415.4 mm at setback 180, so setback 250 would be about 485 mm, roughly 70 mm deeper than E and 103 mm deeper than M. Frozen E is 33 mm deeper and 21 mm taller than M. In J arithmetic, the compact choice gives up about 0.004 in exchange for about 70 mm less depth."
    },
    {
      "id": "C7",
      "result": "verified",
      "title": "Collisions vs projected overlap in images",
      "detail": "In the side views, fingers overlap the keybed front, and the hands overlap the pad and the bridge silhouette in 2D. The three-quarter views (E-6 typing and music) show the hands above the keys and in front of the bridge, with no real intersection. Exported penetration is 0, and the bridge clearance is 40 mm at the white-key target. At E age 3 typing, the wrist-to-pad gap is 0 mm on both sides: the static wrist touches the pad top. That is contact, not penetration, and is consistent with resting the palm on the pad during pauses. It leaves no hover margin for the smallest reference child."
    },
    {
      "id": "C8",
      "result": "native evidence referenced (I did not run the UI browser myself)",
      "title": "Comparison graphs",
      "detail": "native-comparison-ui.json covers desktop 1440\u00d71080 and mobile 390\u00d7844 with default model M and a fixed camera through compare/person/age/activity changes, with 0 errors. Points: typing 40 (M, D, E, W), music 20 (M, E), weighted 20 (M, E). Of the comparison data, 32 weighted rows pass the arithmetic check. M and D raw typing metrics are identical, and D/W music is absent. The graph's raw reach metric is labelled horizontal. A duplicate 0.1 tick on the loss axis was sent to root for polish; it is not a blocker."
    }
  ],
  "realBenefitAndLimitations": {
    "benefit_raw": [
      "Table lift over the neutral-hand construction (the lift beyond the neutral-hand table that auto furniture accepts to preserve leg clearance): typing 111\u2013120 \u2192 83\u201391 mm, music 53\u201359 \u2192 29\u201335 mm, at every age. Easier to read in arm terms: typing forearm slope is 3.8\u20138.2\u00b0 lower, music forearm slope is 4.6\u20136.6\u00b0, and age-3 typing wrist goes from \u22128.5\u00b0 to \u22122.0\u00b0.",
      "Forearm rise: typing reduced 3.8\u20138.2\u00b0 (still +12.9\u2026+19.5\u00b0); music +4.6\u2026+6.6\u00b0.",
      "Shoulder flexion, typing: age 3 31.6\u219225.0\u00b0, age 6 28.6\u219215.7\u00b0. Music 3.6\u20135.5\u00b0.",
      "Age 3 typing wrist: \u22128.5\u00b0 \u2192 \u22122.0\u00b0.",
      "Screen incidence at ages 6\u201312: 8\u201318\u00b0 \u2192 0.3\u20138\u00b0. Gaze at age 12: 28.8 \u2192 26.2\u00b0 (typing).",
      "Front of the music bay is clean (ports moved to the sides), with no visible clutter."
    ],
    "limitations_raw": [
      "Music-only subgroup (final optimizer): E loss exceeds M at nominal ages 3\u20137 (e.g. age 3: 0.0242 vs 0.0041; age 4: 0.0086 vs 0.0010), mainly through viewing (shallow gaze, screen top above the eye). Raw music arm values still improve (forearm slope, shoulder flexion, table lift). The 70/30 aggregate and typing improve at every age; music-only does not.",
      "Typing still needs 83\u201391 mm of table lift over the neutral-hand construction to keep leg clearance, and typing forearms still rise 12.9\u201319.5\u00b0. The app shows 'Keys sit high for relaxed arms'. The fixed GK64X/LPK25 stack and knee-limited auto furniture cap the gain.",
      "Screen top is ABOVE the age-3 reference eye: 8.7 mm in typing and 5.7 mm in music. M is 12.7 and 16.7 mm below. Age-3 gaze is shallow (10.3\u00b0 / 9.5\u00b0) and incidence is worse than M at age 3 (7.7\u20138.5\u00b0 vs 2.4\u20133.7\u00b0). This is reported against OSHA's ADULT top-at-or-below-eye reference. It is not a child hard limit, and no reviewed source gives a numeric pediatric threshold.",
      "Age-12 typing gaze of 26.2\u00b0 remains above the adult 15\u201320\u00b0 reference. A single fixed screen cannot centre both ends of a 3\u201312 range; this is an inherent tradeoff.",
      "Viewing distance at age 6 typing is 19 mm shorter than M. Elsewhere it is within \u00b15 mm.",
      "Static single poses only. No dynamic reach across both octaves, no far-key repositioning, no movement, no fatigue. Body proportions are estimates.",
      "At \u22128% stature (sensitivity), the age-3 screen top is about 35\u201337 mm above the eye and gaze is about 6\u00b0. The smallest young children are where the fixed-screen tradeoff is largest."
    ]
  },
  "resolvedConcerns": [
    {
      "id": "initial V2 (incidence)",
      "status": "resolved",
      "evidence": "The incidence band is in the view subweights (0.25); native audits compute it."
    },
    {
      "id": "initial V3 (correlated viewing)",
      "status": "resolved",
      "evidence": "One view category (0.25), with declared subweights and head tilt down-weighted for its correlation with gaze."
    },
    {
      "id": "initial V4 (discontinuities)",
      "status": "mitigated",
      "evidence": "The chosen point was re-solved, and 26 one-variable neighbours were checked with no hard flags. A continuous hip refinement was not implemented."
    },
    {
      "id": "initial V5 (reach)",
      "status": "primary resolved; secondary pending",
      "evidence": "The primary gate covers the representative target, with reach miss \u2264 5 mm and normalised wrist displacement. Far-key and control reach with repositioning are untested (secondary)."
    },
    {
      "id": "initial V6 / correction 7 (music bridge)",
      "status": "resolved at geometry level",
      "evidence": "Hard bridgeCheck at the white-key target (fingertip/knuckle/palm/wrist radii + 3 mm): 40 mm clearance. Secondary black-key contact with the corrected CAD datum: exposure 25.05 mm (\u2265 23 mm) and capsule clearance conservatively 8.65 mm (\u2265 3 mm). These are not a whole-keyboard fit claim. Donor and finger trials are pending."
    },
    {
      "id": "initial V7 (furniture)",
      "status": "resolved",
      "evidence": "Auto furniture uses the same algorithm for every model."
    },
    {
      "id": "initial V9 (component floors)",
      "status": "partially resolved",
      "evidence": "Deck 62.5 = drop 23 + wheel envelope 30.5 + 9 margin, conservative. The LPK25 reserve is still photo-derived, and its measurement is a production gate."
    },
    {
      "id": "provisional I4 (grille alignment)",
      "status": "resolved by complete evidence on frozen params; no fix requested",
      "evidence": "grille-alignment-final.json was independently re-run on the exact frozen params (245/18/180, rear 92). Screen and grille Y min = 167.19759085844805 mm and Y max = 323.14232783536437 mm, with exactYAlignment = true. /tmp/otterchip-e-final-E-orthographic-front.png visibly confirms it. My provisional observation was a perspective artefact. The earlier grille-alignment.json was for the provisional geometry and is superseded."
    },
    {
      "id": "provisional O4 / I5 (side features)",
      "status": "withdrawn",
      "evidence": "The upper side rectangle is the working volume wheel. Side jacks and the power rocker are recessed near the table behind the keyboard region. The generic USB rectangles have been removed."
    },
    {
      "id": "provisional O3 top strip",
      "status": "clarified",
      "evidence": "The top strip holds stereo microphones. There is no camera."
    },
    {
      "id": "reach presentation",
      "status": "corrected",
      "evidence": "The graph shows horizontal shoulder-joint\u2192wrist in mm/cm, labelled as such. The loss uses wrist displacement from neutral \u00f7 (upper arm + forearm). My chain ratio is descriptive only. Native UI evidence confirms the graph labels raw reach as horizontal."
    },
    {
      "id": "stature labels",
      "status": "corrected",
      "evidence": "The optimizer's robustness is \u00b18% stature SENSITIVITY, not percentiles. My earlier P5/P95 values came from actually fetched CDC statage.csv columns and are not used for the optimizer."
    },
    {
      "id": "final optimizer path",
      "status": "adopted",
      "evidence": "results.json uses the corrected nominal-profile path (no rounded stature override for factor 1). Its E weighted J is 0.014797756266410417, and the refreshed native selected-check.json gives the identical value. The earlier 0.0147936 snapshot differed only by a tiny stature-rounding path. The final optimizer values are canonical here. Site geometry scores and the native 4-age values were valid for their captured solutions."
    }
  ],
  "visualAssessmentAtFinalProportions": {
    "observations": [
      "E-body vs M-body: E reads as a clean, fluid fixed body. It has a lower keyboard, a taller and slightly steeper screen, and a deeper concave sweep. Compared with the provisional 250/20/159, the compact 245/18/180 looks less monumental. The upper mass is better balanced over the deck, though still taller than M.",
      "The front of the music bay is uncluttered: no front jack or switch plates, only the key opening and the small pull tabs. This is cleaner than M's front ports.",
      "The broad cream sweep between the top key row and the screen is the dominant surface. Given the stated preference for a clean shell, it reads as intentional negative space and does not need features added.",
      "The volume wheel at the upper side and the recessed low side plate are discreet and read well in the side views. The stereo mic strip is unobtrusive.",
      "In the 6-year-old three-quarter view, the pad reads flush and calm (7 mm rise)."
    ],
    "recommendations": {
      "required": "None for appearance.",
      "optionalNoFunctionalLoss": [
        "Keep one corner-radius family. Round 4 and grille corner 4 already match; carry the same radius to the recessed side plates and pull-tab pockets in CAD.",
        "Choose a low-gloss or matte finish for the large cream sweep. Large printed surfaces show layer lines, seams and fingerprints. Confirm achievable surface quality with the chosen print process and vendor; this is not a styling change.",
        "No added grooves, logos or light bars are needed."
      ],
      "requiresObjectiveReevaluation": "Any change to height, tilt, setback, deck, drop, underlap, pad rise or gap, or front-lip geometry. The key and screen geometry is the optimized core and should not change casually."
    }
  },
  "chartPresentationNotes": [
    "Native UI evidence shows the comparison graphs meet the earlier presentation requirements: typing for all 4 models, 70/30 only for M and E, D/W music absent, horizontal reach labelled. A duplicate 0.1 loss-axis tick is minor polish.",
    "Rename or explain the ambiguous UI metric keyAboveNeutralMm wherever it is shown, as \"table lift over neutral-hand construction (for leg clearance)\", and prefer forearm slope and wrist angle for readers.",
    "Label J as a designer heuristic, not a fitness score, and show no percent improvement. Label \u00b18% as sensitivity, not percentiles."
  ],
  "blockers": {
    "geometryOrVisual": [],
    "note": "No blocking geometry or visual error. The age-3 screen-top/gaze tradeoff is a disclosed design tradeoff, not a blocker."
  },
  "pendingGates": {
    "mathematicalGeometryDone": [
      "App-solver poses at ages 3\u201312 for typing and music, with hard flags 0 (optimizer: 60 cases; my recheck: 16 cases).",
      "Native equivalence: 113 checks, 0 failures; browser geometry equals the optimizer adapter.",
      "White-key-target bridge check (40 mm). Secondary black-key contact: CAD exposure 25.05 mm, capsule clearance \u2265 8.65 mm. Arm penetration 0.",
      "One-variable neighbourhood robustness (26 variants).",
      "Final grille alignment exact on the frozen params.",
      "Comparison UI and data audit (native)."
    ],
    "pendingCadAndManufacturing": [
      "Build 9 CAD/export checks pass (QA 05:24:51Z; native audit 746 checks, 0 failures; see review-manufacturing.json). The stale supervisor metadata wrappers are root packaging metadata only.",
      "Coupon measurements, opened-donor and display measurements, and the black-key hand test with a mock bridge. The full shell is conditional on these and on the acoustic prototype.",
      "3D-print process checks remain: material for a child product, heat near the electronics, layer-line strength at clips and load paths, and vendor tolerance and finish."
    ],
    "pendingPhysicalValidation": [
      "Child sitting and loading: leaning or pressing on the deck, pad or music cover.",
      "Tip stability: E is 21 mm taller and 33 mm deeper than M, with a tall screen mass.",
      "Drop and impact testing, small parts, pinch points, edge sharpness, and material/finish safety for the applicable product standards (to be determined; not assessed here).",
      "Thermal behaviour (display, SoC, amplifier inside a closed shell), electrical safety, power supply and EMC.",
      "Acoustics: speaker mounting, grille open area and mic placement.",
      "Real-user fit trials with children across ages, furniture and actual typing/playing movement. The static model cannot replace them."
    ]
  },
  "knownLimitations": [
    "Body proportions are age-interpolated historical estimates, with CDC-midpoint stature. No clinical optimum is claimed.",
    "Only 17 images were read. The final three-quarter views cover age 6 only.",
    "This is not a mechanical CAD review and does not certify production readiness.",
    "My earlier per-activity score recomputation is approximate. The canonical scores are from the native core.score cross-check.",
    "Comparison UI checks are native evidence; I did not run them."
  ],
  "revision": "r2 (2026-11-06): bounded factual correction pass. Changed: canonical scores, neutral-hand terminology, black-key bridge datum, final grille evidence, native UI comparison evidence, 3D-print gates, scope. Independent judgments otherwise unchanged. | r3 (2026-11-06): added the final optimizer results (optimization/results.json, generatedAt 2026-11-06T05:03:35.700Z, corrected nominal-profile path with no rounded stature override at factor 1), the music-only subgroup tradeoff, and the \u00b18% screen-top values; earlier 4-age analysis marked approximate. | r4 (2026-11-06 ~10:30Z): final scores aligned to results.json/selected-check (J 0.014797756266410417 vs M 0.052349737283120416). Manufacturing status now points to the final disposition in review-manufacturing.json (build 9 QA 05:24:51Z).",
  "scope": "Final-geometry review of the frozen compact E: rendered appearance, the app solver poses, and the optimizer objective. The mechanical CAD/QA final disposition (build 9, QA 2026-11-06T05:24:51Z) is in review-manufacturing.json. This is not a production certification.",
  "finalOptimizerResults": {
    "source": "optimization/results.json generatedAt 2026-11-06T05:03:35.700Z (optimizer output, not my computation)",
    "weighted7030": {
      "M_J": 0.052349737283120416,
      "E_J": 0.014797756266410417,
      "E_worstAge": 3,
      "note": "E is lower at every age 3\u201312 in the 70/30 per-age values"
    },
    "typingOnly": {
      "M_J": 0.06511310186924327,
      "E_J": 0.01715387415148847,
      "note": "E is lower than M at every age 3\u201312. D equals M (same typing geometry)."
    },
    "musicOnly": {
      "M_J": 0.0225685532488338,
      "E_J": 0.011664260124245915,
      "perAge_M": {
        "3": 0.00414,
        "4": 0.000968,
        "5": 0.001216,
        "6": 0.001435,
        "7": 0.000589,
        "8": 0.002672,
        "9": 0.007681,
        "10": 0.016119,
        "11": 0.030303,
        "12": 0.04966
      },
      "perAge_E": {
        "3": 0.024163,
        "4": 0.008647,
        "5": 0.001718,
        "6": 0.001537,
        "7": 0.000649,
        "8": 0.000713,
        "9": 0.000748,
        "10": 0.002293,
        "11": 0.006083,
        "12": 0.012598
      },
      "subgroupTradeoff": "At nominal stature, E music-only loss is WORSE than M at ages 3\u20137 (largest at ages 3\u20134, driven by viewing: shallow gaze and the screen top above the eye) and better at ages 8\u201312. The aggregate music J is lower for E because ages 9\u201312 dominate. Do not claim that every activity improves at every age."
    },
    "statureSensitivity_pm8pct": "E age 3 at 0.92 stature: screen top 36.8 mm ABOVE the eye in typing (34.8 mm in music), gaze 6.1\u00b0 / 5.9\u00b0, incidence 11.9\u00b0 / 12.1\u00b0. At 1.08 stature it is 20.3 / 23.3 mm below the eye. This is sensitivity, not percentiles.",
    "note": "J is a designer heuristic; no percent improvement or clinical claim."
  },
  "nativeReleaseWrapperReconciliation": {
    "reviewer": "Native release coordinator, separate from actual Claude review",
    "generatedUtc": "2026-11-06T10:28:35.496803+00:00",
    "qaGenerated": "2026-11-06T05:24:51.505517+00:00",
    "qaSHA256": "66c1f2dc42807a5d19fd7b1d60ee5fbb6cbf5b906fc4a96097717ff638bc08f9",
    "assemblyManifestGenerated": "2026-11-06T10:27:00.369923+00:00",
    "assemblyManifestSHA256": "dc1a4b827a088ee36a84db7e5faa753ece51843e4a9553a7b38a1d6f7be032b9",
    "assemblyPartCount": 83,
    "allAssemblyMeshHashesMatch": true,
    "nativeExportAuditGenerated": "2026-11-06T10:25:48.995958+00:00",
    "nativeExportAuditSHA256": "9772376714f84fc84d1c7f96296410a17316480f3f1ae5e37dd2101fa2f58b4b",
    "currentWrappersPass": true,
    "note": "The actual Opus report describes supervisor wrappers as stale at its initial read. They have since been refreshed by the mechanical handoff coordinator and independently hash-checked here. Vendor detailed reference shapes are deliberately excluded from the public assembly; fabricated geometry is unchanged. No physical validation performed."
  }
}
