Recommendations — second handheld UX review
What works in the reviewed evidence
The ordinary-reboot Home image shows a complete portrait layout with a useful separation between the app grid, time/date, page indicator, dock and system gesture handle. Authored dark drawer rendering has recognizable icons, concise labels and search without heavy row outlines. The new loaded Settings captures show consistent filled controls in both dark and light appearances. These are pixel-level observations in the evidence classes named in baseline, not blanket glass acceptance.
What remains unfinished
The output-picker capture has an optional visual-density question, not a demonstrated blocked action. The current Home clock’s heavy outline/shadow draws strong attention against the chosen wallpaper; any quieter alternative needs user review and must preserve their background choices. Current-source card presentation and keyboard-visible reachability are not in the new integrated capture set. Drawer overflow/search and empty/loading/stale/error/ denied states need their own recheck; missing states are not reported as product defects here. The three unchanged real-finger checks accepted on the exact candidate are recorded in operator-navigation.json; they were not repeated after ordinary reboot.
webOS-inspired gap
The useful reference gap is continuity and discoverable recovery: direct interaction should make the active object and resulting state apparent, while common app and escape routes remain visible. Palm webOS’s card/launcher structure and webOS OSE’s app bar/launchpad are contextual comparisons, not a template. K230’s separate drawer, Overview, Home and fixed shell controls should remain coherent on their own terms. No blur, stack, tile grid, extra chrome or copied iconography is recommended by this round.
Shared visual and interaction direction
Reuse the existing portrait interaction contract as the shared source of truth: 568×1232 logical portrait composition; 8/16/24px micro/control/content rhythm; 56px minimum primary target; distinct title, state and value roles; text/state cues that do not rely on color alone; and visible fallback for gesture routes. Keep the authored Omarchy palette, wallpaper and icon inputs intact. Current settings/home/drawer work already shows that quieter filled surfaces and icon-led discovery fit those rules; the picker should reposition within them instead of adding outlines or a new button. Card transitions should preserve the same object/context during drag and preserve close/refusal/recovery semantics. Under reduced motion, settle without removing the interaction or its visible result.
The performance boundary is similarly restrained: prefer layout/paint changes that reuse current primitives. Do not add blur, shadow layers, previews, or per-frame animation as a polish shortcut. Theme chooser evidence records that pending thumbnail work once forced continuous redraws on the single-core board; keep redraws tied to real state changes and any loading pulse bounded to actual pending work. This review contains no new CPU/memory measurement, so any renderer-heavy successor must record its own budget and rollback path.
Next three priorities
| Order | Priority | Existing route and dependency |
|---|---|---|
| 1 | Resolve reported picker cadence/release-settling concerns against a current matched baseline/candidate and complete the already-scoped motion work. | the-shell-swaps-themes-without-a-python-stall tasks 11.3, 17.3, 18.2 and 18.4, plus the retained wider 13/15 gates. The persistent system now has the combined source, but the old pair used a different Rust source; no fresh board timing or finger evidence is inferred here. |
| 2 | Review current card/keyboard motion and recovery as one focused integrated candidate session, reusing the exact Home/Overview/Apps/Terminal interactions already operator-accepted. | This change tasks 3.1–3.2 plus the existing card/keyboard owners. New optical/finger capture only for changed or still-unknown motion/reach claims; keep accepted behavior and avoid a redundant broad capture. |
| 3 | Reconcile theme-authored token/control/icon consistency across Home, drawer, Settings and deck without changing the user’s selected Omarchy palettes, wallpaper or upstream icons. | the-shell-loads-omarchy-themes remaining consumer/icon coverage and the-handheld-presents-a-coherent-shell task 6.4; preserve each owner’s board and real-finger gates. Start with matched host captures/source review; no new borders or permanent controls. |
Priorities 2 and 3 are review/acceptance work, not claims that the implementation is defective. The clock and picker items remain optional visual questions, not selected work. These recommendations do not alter Omarchy themes, colors, backgrounds or icons and do not duplicate GPU, card, recovery, or launcher implementation proposals. This list and the findings reflect the coordinator critique received 2026-10-01; group 3’s still-needed focused candidate review and task 4.2 publication remain open.
Current candidate update — 2026-10-09
The candidate recheck now supplies current Overview, real 19-entry Apps and loaded Settings evidence from the identified normal mainline hotplug runtime, reusing the matching operator Home/Apps/Settings acceptance. This completes the capture/identity/recheck portion of priority 2. The operator also confirms Terminal keyboard, two-app switching and video functional journeys. Optical tracking, exceptional-state recovery, accessibility and the other runtime owners remain open. The same three priority owners and optional clock/picker decisions stand.