--- name: parked-nv-overhaul-design-audit description: "LIVE owner rulings that future design audits must NOT re-flag (per-instrument variation is UNIQUE ON PURPOSE; the SETTINGS account/saves panels are intentional), plus the owner-committed end-of-overhaul design gate and the 'light-frame is not done' quality bar." metadata: node_type: project type: project originSessionId: parked-nv-overhaul-design-audit --- # ⛔ STANDING RULINGS — DO NOT RE-FLAG THESE Owner rulings from the New Vegas design audit (2026-07-08). These are **intentional design, not defects.** Any future audit that reports them is reporting a non-finding. - ⭐ **Per-instrument number-setting variation is INTENTIONAL.** The audit's "six ways to set a number" theme (SPECIAL steppers+drags, skills drag, etc.) was ruled by the owner: these gestures are **"UNIQUE ON PURPOSE"** — the custom-1-of-1-per-screen philosophy ([[design-overhaul-direction]]). **"Do NOT homogenize them."** Other fixes may proceed only *"if they don't affect that"* — i.e. do not touch or standardize the number-setting gestures. - **The SETTINGS account and saves panels are LEFT AS-IS ON PURPOSE.** The owner deliberately left them in their existing style. Do not redesign them; do not report them as legacy-dressed. ⚠ **Note the live tension with the UI-consistency brief** ([[parked-ui-consistency-audit]]), which wants build-failing guards when a panel deviates from a shared pattern. Run as written, that guard **violates the ruling above.** Mechanical consistency enforcement, if it happens at all, must be narrow (density, contrast, tap targets) and must never police per-instrument interaction gestures. # ⚠ THE MEASUREMENT INCIDENT — why audit tooling on this app is not trusted by default One audit "must-fix" (a claim that mobile nav failed to switch panels, backed by a huge measured page height) was a **FALSE ALARM.** Live testing on a local server and on deployed staging showed the nav had always worked, and the **owner confirmed: "it has always switched cleanly" on their phone.** **Root cause: the audit's screenshot tooling hit a RASTERIZATION STALL on this animation-heavy app** (continuous rAF loops) and read the DOM *before* CSS/JS applied — so every panel looked visible at once. ⇒ **LESSON, still live: automated design-audit measurements of this app are UNRELIABLE unless animations are frozen first.** Any future audit of this app needs a reliable way to freeze animation and *then* measure. A measured layout number from a stalled render is not evidence. # ⭐ THE OWNER'S QUALITY BAR — "light-frame ≠ done" Owner directive 2026-07-05, after seeing several boards still in near-legacy form under a dressed frame. The hero-three pattern (deep-dress a few, light-frame the rest) is **not acceptable as an end state.** - Every remaining panel needs a **GROUND-UP redesign, not an "update."** In the owner's words, these *"should not simply be 'updated' — they need to be redesigned from the ground up using the same philosophy as the rest."* - A screen is **NOT done** until every board, control, widget, indicator, graph, list, animation and interaction has been reimagined against: **"Is this the coolest, clearest, most immersive RobCo-inspired implementation possible?"** — not merely "does it work?" - The exemplar is the tempo-knob redesign: changed because it communicated its purpose better *while staying unmistakably Fallout*. - A label is not an implementation — a board called a "VU ARRAY" that renders plain number boxes has not been redesigned. It needs to become a real instrument. - Applies to **every** subsystem, not just the one that prompted it. No board left light-framed. # THE END-OF-OVERHAUL DESIGN GATE (owner-committed, not optional) After the overhaul is built — **before it counts as finished, and before the next game's build and release** — run a dedicated, repository-wide review of design + UX + immersion + in-universe authenticity + feature philosophy. **This is not a visual-polish pass.** It examines: - Visual and interaction consistency across every panel. - Does every screen feel like the SAME operating system? Are any screens too visually SIMILAR? Has any screen become GENERIC instead of uniquely RobCo? - Is the modernized-in-universe philosophy consistently followed? Does each subsystem have its own recognizable identity? Does the implementation support IMMERSION, not just functionality? - Opportunities to make interfaces more diegetic — believable as software someone in the Fallout universe would actually use. - **Game-specific identity:** does this feel uniquely like *this* game? Could each supported game carry subtle differences while sharing the RobCo U.O.S. framework, **without** fragmenting the design language? - **For every weak area, propose MULTIPLE improvements, not one.** Be willing to REPLACE an existing design if a stronger Fallout-native solution exists. **Goal, stated by the owner:** NOT absolute uniformity — *a cohesive design language where every subsystem is immediately recognizable as part of RobCo U.O.S., yet every subsystem has its own memorable identity and purpose.* **Roadmap authority granted:** if during the overhaul the implementation ORDER should change to better achieve this vision, **update the roadmap and explain why** — optimize for the strongest final product, not for preserving the current plan. **How to run it:** the Fable(design) → Opus(audit) split; output is a triaged doc with multiple proposals per finding. The diegetic-audit brief is the house standard ([[parked-diegetic-audit]]). ⭐ **PLANNING-FOLDER TIDY — a DELIBERATE, OWNER-AUTHORIZED exception (2026-07-08).** Bundled with this end-of-overhaul cleanup: sweep `planning/` and `planning/mockups/` to clear genuinely obsolete leftovers (superseded mockups, dead plan docs, stale slates, one-off screenshots). ⚠ This is an explicit owner exception to the standing "never touch `planning/`" rule ([[delete-leftovers-rule]]) — it is a **scheduled, one-off tidy, NOT the routine junk sweep**, and it does not generalize. **PRESERVE anything still live** (the active build queue, master plan, anything a later round still needs). Judgment call per file; **when unsure, keep.** `planning/` is gitignored, so no commit is involved. **Deferred into this gate by owner decision:** the evaluation of a diegetic cold-boot onboarding beat — the gate decides whether it is worth building, since it reviews immersion anyway. Ceremony polish work was deferred as non-blocking. --- *Correction 2026-07-20: this file previously carried a release pre-flight checklist, version/cache numbers, per-item queue placement and audit-doc paths. All of that is derivable state and has been removed — current release state lives in `QUEUE.md`, `CHANGELOG.md` + git log, and `APP_VERSION` in source. The rulings, the quality bar, the gate and the measurement incident above are the durable content.* Related: [[design-overhaul-direction]] (custom-1-of-1 + escalate ambition), [[parked-diegetic-audit]] (house standard), [[robco-audit-program]], [[per-game-device-form]], [[delete-leftovers-rule]].