[Hallmark] Rework Home, profile picker and Appearance into a calmer library interface #48

Open
opened 2026-09-18 13:28:54 +02:00 by alexander · 2 comments
Owner

Purpose and relationship to #47

Create a dedicated work item for the deeper Home/profile/Appearance review, previously recorded as a comment on #47. This issue owns that deeper scope and the proposed Hallmark design direction below. #47 retains the first audit's contrast/player/recovery findings; do not duplicate those implementations here.

Primary ownership: UI & Experience. Status: design recommendations and source findings, pending rendered validation. Other Settings tabs are explicitly excluded.

Hallmark recommendation: a quiet personal library

The design should feel like a considered audiobook collection: covers provide personality, typography explains the content, and controls stay calm. “Less AI-like” is a design judgment, not a claim about authorship. The repeated glass containers, always-moving labels and similar visual weight across unrelated actions are the main patterns to reconsider.

This is a proposed visual/interaction direction, not an implementation or a replacement of the established Material 3 architecture. Preserve routes, repository/state ownership, accessibility, author-before-title shelf ordering and the user's appearance choices. Other Settings tabs remain outside scope.

1. Give each Home section a purpose-specific composition

  • Continue listening: propose one wider leading resume card with a readable title, chapter/remaining-time context when available, and an unmistakable Resume action. Additional in-progress books remain a compact shelf. Derive the lead item from existing product state; do not invent a competing remembered-book policy.
  • Recently added and Discover: retain compact cover-led shelves with quiet labels.
  • Continue series: show a series identity, completed-book count and named next book; keep it visually distinct from a single-book card.
  • Reserve repeated equal-width cards for content that benefits from comparison. Do not vary layouts randomly or add oversized marketing headings.

2. Use glass selectively

  • Keep translucent treatment on floating navigation and the mini-player, where it communicates layering.
  • Propose calmer, opaque or lightly toned text surfaces for the default book/collection browsing presentation. Avoid an outline, blur and separate rounded panel around every small group of text.
  • Retain the user's glass/background options; evaluate a quieter default and preserve readable scrims for optional artwork themes. Reuse existing surface roles and contrast-tested pairs instead of introducing a second palette.
  • Do not add gradients or parallax solely to make glass noticeable.

3. Make the flat list a reading surface

  • Use compact rows with consistently sized smaller covers, strong title hierarchy, secondary author/series metadata and one concise progress line.
  • Prefer subtle separators or spacing to a procession of equally prominent glass cards.
  • Keep row height adaptive, aligned metadata, and a visible active filter. Use a labelled Sort control to separate ordering from membership filtering.
  • Keep details as the general row action; preserve the existing deliberate distinction between browsing and direct playback.

4. Retain meaningful collection artwork, reduce ornamental repetition

  • Keep series stacks, real author portraits/cached-cover fallbacks, and genre mosaics.
  • Scale them down on narrow screens so names and counts dominate over decorative cover composition.
  • Consider removing repeated Person/Category badges where the selected tab already establishes the type. Keep them only where they clarify mixed content.
  • Use the same spacing/type system across tabs; do not force identical card layouts onto books, people and genres.

5. Keep navigation stable and labels readable

  • Preserve Books / Series / Authors / Genres and swipe navigation.
  • Use one clear selected indicator, persistent readable labels, and restrained transition timing.
  • Adapt to window width and translated labels; avoid decorative bounce or motion competing with scrolling content.
  • Keep tap alternatives for every gesture.

6. Treat the profile picker as an account chooser

  • Give identity the first line: display name, username when needed, server, then concise role/security status.
  • Preserve a clear Current badge and put maintenance actions in a separate account menu.
  • Use a focused Switching state rather than disabling the entire surface without explanation.
  • Keep destructive confirmation copy and permissions unchanged; reserve bottom clearance for playback and constrain width on tablets.

7. Turn Appearance into a small, understandable preview tool

  • Keep the current Theme → Colour → Glass → Language organization.
  • Add one live sample built with actual BookWave components: book text, a control and the selected surface. No fake phone/browser frame.
  • Show named theme/color choices, a clear selected mark and current source (explicit color, wallpaper, follow accent).
  • Keep color selection open for comparison; use progressive disclosure for detailed glass adjustment.
  • Label effects in product terms, such as “Navigation and player tint,” and explain platform fallback.
  • Add Restore appearance defaults with a clear scope. Keep language and profile/security settings outside that reset unless explicitly included and communicated.
  • Fix the pack/light-dark preview discrepancy before polishing the preview.

8. Establish a restrained component rule set

Use the existing Material 3 typography, semantic colors and icon family. Document a small spacing scale, distinct roles for shelf covers versus list thumbnails, a consistent corner hierarchy, and restrained motion. Default to static metadata; any marquee should be focused, deliberate and reduced-motion aware. Use color for selection/status with a non-color cue. Do not add decorative badges, new fonts or arbitrary asymmetry to manufacture uniqueness.

Delivery slices and acceptance

  1. Fix adaptive heights, player clearance, preview resolution and accessible progress meaning.
  2. Prototype the leading Continue card, quiet flat rows, narrower collection artwork and simpler profile actions using real representative metadata.
  3. Compare those prototypes with the existing presentation at narrow/large-text widths before implementation.
  4. Refine Appearance preview, named color sources, selective glass and motion.
  5. Run Hallmark's final hierarchy/restraint/state/contrast review and correct material issues.

Keep current routes/components where practical; change the smallest visual layer needed. No rebuild or deletion plan is authorized by this issue. Preserve the canonical roadmap's correctness sequencing.

Accept only after rendered comparison at 320/375/414/768 dp, large-text scales through 2.0, long labels, missing artwork, active playback, dark/light/AMOLED/background packs, TalkBack and reduced motion. Measure glass contrast against effective backgrounds and test older-Android fallback. A source review does not prove rendering.

The intended result is recognizable as BookWave through its library and listening workflow, with fewer competing surfaces and clearer action priorities.


Full source-based findings

BookWave: Home, profile picker and Appearance design review

Reviewed 18 September 2026 against main at 8f0ff5926687b3373703ee972dbb1392a0792d63.

Scope and evidence

This second Hallmark audit covers Home shelf cards; Books, Series, Authors and Genres tabs; the flat book list; profile picker; and Settings → Appearance. Other Settings tab contents are excluded because they have not yet been improved.

All implementation evidence was read from Forgejo through prox2-mcp-gateway. The original GitHub repository was used only to locate filenames because the gateway exposes no tree-listing tool. No application files were changed. This is a source-based review, not a rendered-device acceptance pass. Layout risks below need reproduction at the stated sizes; aesthetic recommendations are identified as recommendations rather than observed rendering failures.

The first audit is tracked in issue #47. Its contrast and mini-player findings remain relevant and are not counted again here.

Design direction to preserve

Keep BookWave cover-led and native. Home shelves support recognition and resumption; the flat list supports scanning and finding; collection tabs organize the library; the profile picker should prioritize account identity and switching; Appearance should show the effect of a choice before requiring the user to interpret settings terminology.

Good existing decisions:

  • Author portraits have cached-cover fallbacks; series use stacks and genres use mosaics. These are meaningful differences, not decorative variation.
  • Empty Home shelves disappear, and play/resume has its own labelled control alongside the details action.
  • Series/group cards already use minimum heights rather than fixed heights.
  • The profile picker has an active badge, server identity, passcode information and reauthentication copy.
  • Appearance has coherent Theme → Colour → Glass → Language groups, inline disclosure, 48 dp swatch cells and a named wallpaper-color option.

Keep the established author-before-title shelf order and compact player preferences unless deliberately reconsidered. No wholesale rebrand, replacement navigation system or removal of glass is proposed.

Major findings

1. Flat list: fixed row height conflicts with scalable content

Hallmark lens: responsive layout and typography. feature/browse/BookListUi.kt:83; row height is 132 dp at line 195.

A row must accommodate a two-line title, author, optional series, progress copy, progress bar and 32 dp vertical padding. The series and progress text can wrap, while the outer height cannot grow. This creates a strong clipping/compression risk at larger font scales and with long metadata.

Change: use a minimum height and content-driven expansion. Decouple the cover from full row height so increasing text size does not enlarge the cover and further reduce text width. Preserve equal baseline density at default size.

Acceptance: long title + long series + progress visible at 320/360 dp and 2.0 font scale; no missing lines or overlapping controls. This needs a rendered test, not a source-only pass.

2. Profile picker: the last actions can sit behind the mini-player

Hallmark lens: overlay/layout safety. feature/profiles/ProfileSwitcherScreen.kt:126 and MainActivity.kt:277.

The profile list has only 16 dp content padding. The application draws the mini-player over the navigation graph, and this screen does not consume the player clearance used elsewhere. With enough profiles to fill the screen, the last card or Add profile action can remain under that overlay.

Change: add measured player clearance to the list's bottom content padding. Use a bounded readable width on tablets while touching this layout.

Acceptance: with a loaded book and a long profile list, scroll every final action fully above the player at normal and large text sizes.

3. Appearance: preview colors can disagree with the selected background pack

Hallmark lens: design-system consistency. feature/settings/SettingsScreen.kt:404, MainActivity.kt:112, feature/settings/AppearanceTab.kt:177.

The activity resolves darkness from the selected background pack first. Settings instead supplies appearance.theme.isDark ?: systemIsDark to Appearance. Accent swatches and wallpaper preview then use that value. A dark pack over a light base/system setting can therefore preview the light variant while the app renders the dark one, or vice versa.

Change: use the same resolved appearance value for the application and all preview controls. Keep that resolution in the existing appearance ownership boundary.

Acceptance: dark pack + light system, light pack + dark system, explicit accent and wallpaper accent all show the color actually applied.

4. Home cards: several perpetual marquees compete for attention

Hallmark lens: unnecessary continuous motion. feature/home/HomeShelvesUi.kt:297 and feature/home/HomeShelvesUi.kt:381.

Author, title and series each use an infinite marquee on 160 dp cards. Several visible cards can animate several lines simultaneously, especially at large font sizes. This makes a recognition surface harder to scan.

Change: preserve full-title access while reducing simultaneous movement: allow a larger card/text layout at large font sizes, or animate only the explicitly focused/selected card. Provide a static reduced-motion treatment and keep complete metadata available in details and accessibility text. Treat two-line titles as an option to evaluate, not an automatic reversal of the prior compact-card preference.

Acceptance: a shelf of long metadata remains readable without watching multiple moving lines; reduced motion does not hide the full title.

5. Home cards: progress readings lack explicit meaning and room to grow

Hallmark lens: hierarchy, accessibility and responsive text. feature/home/HomeShelvesUi.kt:340.

Elapsed time, percentage and remaining time share one unweighted row in roughly 136 dp of inner width. They are bare values, and the source explicitly acknowledges that their spoken meaning depends on order. Long durations and larger fonts have little room.

Change: prioritize “time left” visually, or retain all three in an adaptive two-row treatment. Provide a complete card announcement that includes book identity and labels each reading; avoid a child contentDescription that replaces the book title. Keep Play as a separate accessible action.

Acceptance: long-duration values remain intact at 2.0 font scale; TalkBack identifies the book, elapsed and remaining time, and the distinct play action.

6. Collection tabs: artwork uses too much of narrow cards

Hallmark lens: hierarchy and responsive allocation. feature/browse/SeriesListUi.kt:58 and feature/browse/SeriesListUi.kt:115.

Series artwork reserves 140 dp; author/genre artwork reserves 132 dp. At 320 dp with Home's outer margins and the text padding, only about 116–124 dp remains for text before a genre Edit button is added. Cards can grow vertically, but the fixed artwork makes long names unnecessarily narrow.

Change: keep the stack/portrait/mosaic identities, but use smaller bounded artwork on narrow windows or a compact alternate layout at large font sizes. Keep the genre edit action separate from the browse target; move it to a secondary row/menu when width is insufficient.

Acceptance: long series, author and genre names are readable without tall, narrow text columns; edit remains independently reachable.

7. Profile picker: switching and account maintenance have equal emphasis

Hallmark lens: action hierarchy and interaction states. feature/profiles/ProfileSwitcherScreen.kt:211 and feature/profiles/ProfileSwitcherScreen.kt:287.

Inactive cards are tappable and also contain Use, Sign out/Sign in again, and Remove in one row. Routine switching and destructive maintenance occupy the same visual tier. Busy state disables actions but does not show which profile is being switched.

Change: make the profile identity/switch region the primary action, with a separate account menu for sign-out/removal. Keep reauthentication visible when needed and preserve the removal explanation. Display username alongside server when it differs from the display name. Add a targeted “Switching…” state and selected-state semantics; do not introduce another owner for switching policy.

Acceptance: users can identify the active account, choose the correct same-name account on another server, and distinguish switching from removal at large text sizes.

8. Appearance: color choices hide useful identity and close too quickly

Hallmark lens: discoverability and state clarity. feature/settings/AppearanceTab.kt:378 and feature/settings/AppearanceTab.kt:555.

The collapsed color row shows a swatch but not the current color's visible name. Expanded swatches are named only for accessibility, selection uses a stronger ring, and each choice closes the palette. Comparing nearby colors requires repeatedly reopening it. “Follow accent” is a behavioral choice presented largely as another color.

Change: show the selected name/source alongside the swatch, give the selected option a clear checkmark, and keep the palette open during comparison. Present Follow accent as a named source option, like the existing wallpaper option, rather than relying on its resolved hue. Preserve the 48 dp targets.

Acceptance: the chosen color and its source are understandable without distinguishing subtle hues; comparison takes one open interaction.

Minor findings and design refinements

9. Appearance: previews show backgrounds, not the resulting interface

Hallmark lens: specificity and useful feedback. feature/settings/AppearanceTab.kt:95 and feature/settings/AppearanceTab.kt:213.

Theme thumbnails show artwork or flat grounds, while glass controls expose blur values and separate card/system tint switches. They do not demonstrate their combined effect on readable text and controls.

Change: add one small live preview using real shared card/control components, without a fake phone frame. Explain Card tint and Navigation/player tint in user terms; offer an explicit restore-defaults action. Show a platform explanation when blur is unavailable and verify the fallback rather than promising the same glass on every Android version.

10. Home tabs: adapt to available width, not only font scale

Hallmark lens: responsive navigation. feature/home/HomeScreen.kt:739 and feature/home/HomeScreen.kt:861.

The horizontal-to-stacked change depends only on a 1.3 font-scale threshold. Window width and translated label length can cause crowding before that threshold.

Change: choose the arrangement using available width and label fit as well as font scale. Preserve readable labels and tap targets; allow platform-appropriate wrapping when required for accessibility rather than forcing a web-style one-line rule. Validate all four tabs with the mini-player present.

11. Flat list: filter and sort have the same chip treatment

Hallmark lens: hierarchy and restraint. feature/home/HomeScreen.kt:1070 and feature/browse/BookListUi.kt:202.

Search, focused collection, filter chips and seven sort choices can accumulate in the pinned header. Filters change membership; sorting changes order, but they look like parallel categories.

Change: retain the useful All / Continue / Downloaded filters and consider one labelled “Sort: Last played” control that opens the order choices. Show active collection scope clearly. Validate this as a density improvement rather than hiding the way out of an empty result.

12. Card artwork: failed URLs have a weaker fallback than absent URLs

Hallmark lens: complete image states. feature/browse/BookCover.kt:45.

A missing URL draws the book glyph. A non-null URL uses AsyncImage without an explicit loading/error placeholder; failed requests retain the background box but lose the clear fallback symbol. Author imagery already has a more complete fallback treatment.

Change: use the same stable cover placeholder for loading, missing and failed imagery, optionally distinguishing an actionable failure only where useful. Do not change card geometry when loading finishes.

  1. Correct fixed-height rows, profile overlay clearance and appearance preview resolution.
  2. Improve shelf progress readability, collection width allocation and profile action hierarchy.
  3. Refine palette comparison, previews, tabs, filters and image states.

Keep the canonical roadmap authoritative. This review is a candidate UI backlog, not permission to pull visual work ahead of active playback correctness.

Verify on Android at 320/375/414/768 dp widths and larger tablet windows, font scales 1.0/1.3/2.0, portrait/landscape, long English/Norwegian labels, light/dark/AMOLED/background packs, and reduced motion. Include offline/missing covers, long metadata, signed-out and locked profiles, and active playback. Test TalkBack order, selection, expanded/collapsed state, separate nested actions and touch targets. Verify actual glass/background contrast and pre-Android-12 fallback separately.

Verdict: preserve the visual identity and meaningful card families; improve adaptive geometry and make actions and appearance choices easier to understand.

0 new critical · 8 major · 4 minor.

<!-- Hallmark · pre-emit critique: P5 H5 E4 S5 R5 V4 --> ## Purpose and relationship to #47 Create a dedicated work item for the deeper Home/profile/Appearance review, previously recorded as [a comment on #47](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/issues/47#issuecomment-83). This issue owns that deeper scope and the proposed Hallmark design direction below. #47 retains the first audit's contrast/player/recovery findings; do not duplicate those implementations here. Primary ownership: UI & Experience. Status: design recommendations and source findings, pending rendered validation. **Other Settings tabs are explicitly excluded.** ## Hallmark recommendation: a quiet personal library The design should feel like a considered audiobook collection: covers provide personality, typography explains the content, and controls stay calm. “Less AI-like” is a design judgment, not a claim about authorship. The repeated glass containers, always-moving labels and similar visual weight across unrelated actions are the main patterns to reconsider. This is a proposed visual/interaction direction, not an implementation or a replacement of the established Material 3 architecture. Preserve routes, repository/state ownership, accessibility, author-before-title shelf ordering and the user's appearance choices. Other Settings tabs remain outside scope. ### 1. Give each Home section a purpose-specific composition - Continue listening: propose one wider leading resume card with a readable title, chapter/remaining-time context when available, and an unmistakable Resume action. Additional in-progress books remain a compact shelf. Derive the lead item from existing product state; do not invent a competing remembered-book policy. - Recently added and Discover: retain compact cover-led shelves with quiet labels. - Continue series: show a series identity, completed-book count and named next book; keep it visually distinct from a single-book card. - Reserve repeated equal-width cards for content that benefits from comparison. Do not vary layouts randomly or add oversized marketing headings. ### 2. Use glass selectively - Keep translucent treatment on floating navigation and the mini-player, where it communicates layering. - Propose calmer, opaque or lightly toned text surfaces for the default book/collection browsing presentation. Avoid an outline, blur and separate rounded panel around every small group of text. - Retain the user's glass/background options; evaluate a quieter default and preserve readable scrims for optional artwork themes. Reuse existing surface roles and contrast-tested pairs instead of introducing a second palette. - Do not add gradients or parallax solely to make glass noticeable. ### 3. Make the flat list a reading surface - Use compact rows with consistently sized smaller covers, strong title hierarchy, secondary author/series metadata and one concise progress line. - Prefer subtle separators or spacing to a procession of equally prominent glass cards. - Keep row height adaptive, aligned metadata, and a visible active filter. Use a labelled Sort control to separate ordering from membership filtering. - Keep details as the general row action; preserve the existing deliberate distinction between browsing and direct playback. ### 4. Retain meaningful collection artwork, reduce ornamental repetition - Keep series stacks, real author portraits/cached-cover fallbacks, and genre mosaics. - Scale them down on narrow screens so names and counts dominate over decorative cover composition. - Consider removing repeated Person/Category badges where the selected tab already establishes the type. Keep them only where they clarify mixed content. - Use the same spacing/type system across tabs; do not force identical card layouts onto books, people and genres. ### 5. Keep navigation stable and labels readable - Preserve Books / Series / Authors / Genres and swipe navigation. - Use one clear selected indicator, persistent readable labels, and restrained transition timing. - Adapt to window width and translated labels; avoid decorative bounce or motion competing with scrolling content. - Keep tap alternatives for every gesture. ### 6. Treat the profile picker as an account chooser - Give identity the first line: display name, username when needed, server, then concise role/security status. - Preserve a clear Current badge and put maintenance actions in a separate account menu. - Use a focused Switching state rather than disabling the entire surface without explanation. - Keep destructive confirmation copy and permissions unchanged; reserve bottom clearance for playback and constrain width on tablets. ### 7. Turn Appearance into a small, understandable preview tool - Keep the current Theme → Colour → Glass → Language organization. - Add one live sample built with actual BookWave components: book text, a control and the selected surface. No fake phone/browser frame. - Show named theme/color choices, a clear selected mark and current source (explicit color, wallpaper, follow accent). - Keep color selection open for comparison; use progressive disclosure for detailed glass adjustment. - Label effects in product terms, such as “Navigation and player tint,” and explain platform fallback. - Add Restore appearance defaults with a clear scope. Keep language and profile/security settings outside that reset unless explicitly included and communicated. - Fix the pack/light-dark preview discrepancy before polishing the preview. ### 8. Establish a restrained component rule set Use the existing Material 3 typography, semantic colors and icon family. Document a small spacing scale, distinct roles for shelf covers versus list thumbnails, a consistent corner hierarchy, and restrained motion. Default to static metadata; any marquee should be focused, deliberate and reduced-motion aware. Use color for selection/status with a non-color cue. Do not add decorative badges, new fonts or arbitrary asymmetry to manufacture uniqueness. ## Delivery slices and acceptance 1. Fix adaptive heights, player clearance, preview resolution and accessible progress meaning. 2. Prototype the leading Continue card, quiet flat rows, narrower collection artwork and simpler profile actions using real representative metadata. 3. Compare those prototypes with the existing presentation at narrow/large-text widths before implementation. 4. Refine Appearance preview, named color sources, selective glass and motion. 5. Run Hallmark's final hierarchy/restraint/state/contrast review and correct material issues. Keep current routes/components where practical; change the smallest visual layer needed. No rebuild or deletion plan is authorized by this issue. Preserve the canonical roadmap's correctness sequencing. Accept only after rendered comparison at 320/375/414/768 dp, large-text scales through 2.0, long labels, missing artwork, active playback, dark/light/AMOLED/background packs, TalkBack and reduced motion. Measure glass contrast against effective backgrounds and test older-Android fallback. A source review does not prove rendering. The intended result is recognizable as BookWave through its library and listening workflow, with fewer competing surfaces and clearer action priorities. --- ## Full source-based findings # BookWave: Home, profile picker and Appearance design review Reviewed 18 September 2026 against main at `8f0ff5926687b3373703ee972dbb1392a0792d63`. ## Scope and evidence This second Hallmark audit covers Home shelf cards; Books, Series, Authors and Genres tabs; the flat book list; profile picker; and Settings → Appearance. Other Settings tab contents are excluded because they have not yet been improved. All implementation evidence was read from Forgejo through prox2-mcp-gateway. The original GitHub repository was used only to locate filenames because the gateway exposes no tree-listing tool. No application files were changed. This is a source-based review, not a rendered-device acceptance pass. Layout risks below need reproduction at the stated sizes; aesthetic recommendations are identified as recommendations rather than observed rendering failures. The first audit is tracked in [issue #47](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/issues/47). Its contrast and mini-player findings remain relevant and are not counted again here. ## Design direction to preserve Keep BookWave cover-led and native. Home shelves support recognition and resumption; the flat list supports scanning and finding; collection tabs organize the library; the profile picker should prioritize account identity and switching; Appearance should show the effect of a choice before requiring the user to interpret settings terminology. Good existing decisions: - Author portraits have cached-cover fallbacks; series use stacks and genres use mosaics. These are meaningful differences, not decorative variation. - Empty Home shelves disappear, and play/resume has its own labelled control alongside the details action. - Series/group cards already use minimum heights rather than fixed heights. - The profile picker has an active badge, server identity, passcode information and reauthentication copy. - Appearance has coherent Theme → Colour → Glass → Language groups, inline disclosure, 48 dp swatch cells and a named wallpaper-color option. Keep the established author-before-title shelf order and compact player preferences unless deliberately reconsidered. No wholesale rebrand, replacement navigation system or removal of glass is proposed. ## Major findings ### 1. Flat list: fixed row height conflicts with scalable content **Hallmark lens: responsive layout and typography.** [feature/browse/BookListUi.kt:83](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/browse/BookListUi.kt#L83); row height is 132 dp at line 195. A row must accommodate a two-line title, author, optional series, progress copy, progress bar and 32 dp vertical padding. The series and progress text can wrap, while the outer height cannot grow. This creates a strong clipping/compression risk at larger font scales and with long metadata. **Change:** use a minimum height and content-driven expansion. Decouple the cover from full row height so increasing text size does not enlarge the cover and further reduce text width. Preserve equal baseline density at default size. **Acceptance:** long title + long series + progress visible at 320/360 dp and 2.0 font scale; no missing lines or overlapping controls. This needs a rendered test, not a source-only pass. ### 2. Profile picker: the last actions can sit behind the mini-player **Hallmark lens: overlay/layout safety.** [feature/profiles/ProfileSwitcherScreen.kt:126](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/profiles/ProfileSwitcherScreen.kt#L126) and [MainActivity.kt:277](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/MainActivity.kt#L277). The profile list has only 16 dp content padding. The application draws the mini-player over the navigation graph, and this screen does not consume the player clearance used elsewhere. With enough profiles to fill the screen, the last card or Add profile action can remain under that overlay. **Change:** add measured player clearance to the list's bottom content padding. Use a bounded readable width on tablets while touching this layout. **Acceptance:** with a loaded book and a long profile list, scroll every final action fully above the player at normal and large text sizes. ### 3. Appearance: preview colors can disagree with the selected background pack **Hallmark lens: design-system consistency.** [feature/settings/SettingsScreen.kt:404](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/settings/SettingsScreen.kt#L404), [MainActivity.kt:112](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/MainActivity.kt#L112), [feature/settings/AppearanceTab.kt:177](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/settings/AppearanceTab.kt#L177). The activity resolves darkness from the selected background pack first. Settings instead supplies `appearance.theme.isDark ?: systemIsDark` to Appearance. Accent swatches and wallpaper preview then use that value. A dark pack over a light base/system setting can therefore preview the light variant while the app renders the dark one, or vice versa. **Change:** use the same resolved appearance value for the application and all preview controls. Keep that resolution in the existing appearance ownership boundary. **Acceptance:** dark pack + light system, light pack + dark system, explicit accent and wallpaper accent all show the color actually applied. ### 4. Home cards: several perpetual marquees compete for attention **Hallmark lens: unnecessary continuous motion.** [feature/home/HomeShelvesUi.kt:297](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/home/HomeShelvesUi.kt#L297) and [feature/home/HomeShelvesUi.kt:381](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/home/HomeShelvesUi.kt#L381). Author, title and series each use an infinite marquee on 160 dp cards. Several visible cards can animate several lines simultaneously, especially at large font sizes. This makes a recognition surface harder to scan. **Change:** preserve full-title access while reducing simultaneous movement: allow a larger card/text layout at large font sizes, or animate only the explicitly focused/selected card. Provide a static reduced-motion treatment and keep complete metadata available in details and accessibility text. Treat two-line titles as an option to evaluate, not an automatic reversal of the prior compact-card preference. **Acceptance:** a shelf of long metadata remains readable without watching multiple moving lines; reduced motion does not hide the full title. ### 5. Home cards: progress readings lack explicit meaning and room to grow **Hallmark lens: hierarchy, accessibility and responsive text.** [feature/home/HomeShelvesUi.kt:340](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/home/HomeShelvesUi.kt#L340). Elapsed time, percentage and remaining time share one unweighted row in roughly 136 dp of inner width. They are bare values, and the source explicitly acknowledges that their spoken meaning depends on order. Long durations and larger fonts have little room. **Change:** prioritize “time left” visually, or retain all three in an adaptive two-row treatment. Provide a complete card announcement that includes book identity and labels each reading; avoid a child contentDescription that replaces the book title. Keep Play as a separate accessible action. **Acceptance:** long-duration values remain intact at 2.0 font scale; TalkBack identifies the book, elapsed and remaining time, and the distinct play action. ### 6. Collection tabs: artwork uses too much of narrow cards **Hallmark lens: hierarchy and responsive allocation.** [feature/browse/SeriesListUi.kt:58](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/browse/SeriesListUi.kt#L58) and [feature/browse/SeriesListUi.kt:115](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/browse/SeriesListUi.kt#L115). Series artwork reserves 140 dp; author/genre artwork reserves 132 dp. At 320 dp with Home's outer margins and the text padding, only about 116–124 dp remains for text before a genre Edit button is added. Cards can grow vertically, but the fixed artwork makes long names unnecessarily narrow. **Change:** keep the stack/portrait/mosaic identities, but use smaller bounded artwork on narrow windows or a compact alternate layout at large font sizes. Keep the genre edit action separate from the browse target; move it to a secondary row/menu when width is insufficient. **Acceptance:** long series, author and genre names are readable without tall, narrow text columns; edit remains independently reachable. ### 7. Profile picker: switching and account maintenance have equal emphasis **Hallmark lens: action hierarchy and interaction states.** [feature/profiles/ProfileSwitcherScreen.kt:211](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/profiles/ProfileSwitcherScreen.kt#L211) and [feature/profiles/ProfileSwitcherScreen.kt:287](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/profiles/ProfileSwitcherScreen.kt#L287). Inactive cards are tappable and also contain Use, Sign out/Sign in again, and Remove in one row. Routine switching and destructive maintenance occupy the same visual tier. Busy state disables actions but does not show which profile is being switched. **Change:** make the profile identity/switch region the primary action, with a separate account menu for sign-out/removal. Keep reauthentication visible when needed and preserve the removal explanation. Display username alongside server when it differs from the display name. Add a targeted “Switching…” state and selected-state semantics; do not introduce another owner for switching policy. **Acceptance:** users can identify the active account, choose the correct same-name account on another server, and distinguish switching from removal at large text sizes. ### 8. Appearance: color choices hide useful identity and close too quickly **Hallmark lens: discoverability and state clarity.** [feature/settings/AppearanceTab.kt:378](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/settings/AppearanceTab.kt#L378) and [feature/settings/AppearanceTab.kt:555](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/settings/AppearanceTab.kt#L555). The collapsed color row shows a swatch but not the current color's visible name. Expanded swatches are named only for accessibility, selection uses a stronger ring, and each choice closes the palette. Comparing nearby colors requires repeatedly reopening it. “Follow accent” is a behavioral choice presented largely as another color. **Change:** show the selected name/source alongside the swatch, give the selected option a clear checkmark, and keep the palette open during comparison. Present Follow accent as a named source option, like the existing wallpaper option, rather than relying on its resolved hue. Preserve the 48 dp targets. **Acceptance:** the chosen color and its source are understandable without distinguishing subtle hues; comparison takes one open interaction. ## Minor findings and design refinements ### 9. Appearance: previews show backgrounds, not the resulting interface **Hallmark lens: specificity and useful feedback.** [feature/settings/AppearanceTab.kt:95](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/settings/AppearanceTab.kt#L95) and [feature/settings/AppearanceTab.kt:213](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/settings/AppearanceTab.kt#L213). Theme thumbnails show artwork or flat grounds, while glass controls expose blur values and separate card/system tint switches. They do not demonstrate their combined effect on readable text and controls. **Change:** add one small live preview using real shared card/control components, without a fake phone frame. Explain Card tint and Navigation/player tint in user terms; offer an explicit restore-defaults action. Show a platform explanation when blur is unavailable and verify the fallback rather than promising the same glass on every Android version. ### 10. Home tabs: adapt to available width, not only font scale **Hallmark lens: responsive navigation.** [feature/home/HomeScreen.kt:739](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/home/HomeScreen.kt#L739) and [feature/home/HomeScreen.kt:861](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/home/HomeScreen.kt#L861). The horizontal-to-stacked change depends only on a 1.3 font-scale threshold. Window width and translated label length can cause crowding before that threshold. **Change:** choose the arrangement using available width and label fit as well as font scale. Preserve readable labels and tap targets; allow platform-appropriate wrapping when required for accessibility rather than forcing a web-style one-line rule. Validate all four tabs with the mini-player present. ### 11. Flat list: filter and sort have the same chip treatment **Hallmark lens: hierarchy and restraint.** [feature/home/HomeScreen.kt:1070](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/home/HomeScreen.kt#L1070) and [feature/browse/BookListUi.kt:202](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/browse/BookListUi.kt#L202). Search, focused collection, filter chips and seven sort choices can accumulate in the pinned header. Filters change membership; sorting changes order, but they look like parallel categories. **Change:** retain the useful All / Continue / Downloaded filters and consider one labelled “Sort: Last played” control that opens the order choices. Show active collection scope clearly. Validate this as a density improvement rather than hiding the way out of an empty result. ### 12. Card artwork: failed URLs have a weaker fallback than absent URLs **Hallmark lens: complete image states.** [feature/browse/BookCover.kt:45](https://forgejo.homebord.org/alexander/Audiobookshelf-Manager-Claude/src/commit/8f0ff5926687b3373703ee972dbb1392a0792d63/app/src/main/kotlin/com/example/shelfplayer/feature/browse/BookCover.kt#L45). A missing URL draws the book glyph. A non-null URL uses AsyncImage without an explicit loading/error placeholder; failed requests retain the background box but lose the clear fallback symbol. Author imagery already has a more complete fallback treatment. **Change:** use the same stable cover placeholder for loading, missing and failed imagery, optionally distinguishing an actionable failure only where useful. Do not change card geometry when loading finishes. ## Recommended sequence and verification 1. Correct fixed-height rows, profile overlay clearance and appearance preview resolution. 2. Improve shelf progress readability, collection width allocation and profile action hierarchy. 3. Refine palette comparison, previews, tabs, filters and image states. Keep the canonical roadmap authoritative. This review is a candidate UI backlog, not permission to pull visual work ahead of active playback correctness. Verify on Android at 320/375/414/768 dp widths and larger tablet windows, font scales 1.0/1.3/2.0, portrait/landscape, long English/Norwegian labels, light/dark/AMOLED/background packs, and reduced motion. Include offline/missing covers, long metadata, signed-out and locked profiles, and active playback. Test TalkBack order, selection, expanded/collapsed state, separate nested actions and touch targets. Verify actual glass/background contrast and pre-Android-12 fallback separately. **Verdict:** preserve the visual identity and meaningful card families; improve adaptive geometry and make actions and appearance choices easier to understand. **0 new critical · 8 major · 4 minor.**
Author
Owner

Expanded two design recommendations into dedicated implementation issues:

  • #49 — Give Continue listening a clear leading Resume card. Covers composition, existing remembered-book/playback ownership, responsive behavior and acceptance.
  • #50 — Rework the flat book list into readable adaptive rows. Covers thumbnail/text hierarchy, distinct filter/sort controls, shared component boundaries and acceptance.

These are child work items of this design direction; keep the canonical roadmap's sequencing. Collection artwork, profile-picker hierarchy and Appearance preview remain recommendations here.

Expanded two design recommendations into dedicated implementation issues: - #49 — Give Continue listening a clear leading Resume card. Covers composition, existing remembered-book/playback ownership, responsive behavior and acceptance. - #50 — Rework the flat book list into readable adaptive rows. Covers thumbnail/text hierarchy, distinct filter/sort controls, shared component boundaries and acceptance. These are child work items of this design direction; keep the canonical roadmap's sequencing. Collection artwork, profile-picker hierarchy and Appearance preview remain recommendations here.
Author
Owner

2026-09-27 accepted refinement / child split

The owner accepted the Hallmark direction. Focused child issues now make the approved work independently implementable:

  • #49 — leading Continue listening / Resume composition.
  • #50 — flat-list adaptive rows and filter/sort hierarchy.
  • #95 — Home toolbar density, shelf-role differentiation and calmer shelf motion.
  • #98 — profile switcher as an identity-first chooser with maintenance actions demoted.
  • #99 — broader Settings shell/navigation and diagnostics hierarchy.

#48 remains the parent design direction and Appearance-focused staging area. #87 separately owns Home card roundness/border/background preferences.

## 2026-09-27 accepted refinement / child split The owner accepted the Hallmark direction. Focused child issues now make the approved work independently implementable: - #49 — leading Continue listening / Resume composition. - #50 — flat-list adaptive rows and filter/sort hierarchy. - #95 — Home toolbar density, shelf-role differentiation and calmer shelf motion. - #98 — profile switcher as an identity-first chooser with maintenance actions demoted. - #99 — broader Settings shell/navigation and diagnostics hierarchy. #48 remains the parent design direction and Appearance-focused staging area. #87 separately owns Home card roundness/border/background preferences.
Sign in to join this conversation.
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
alexander/Audiobookshelf-Manager-Claude#48
No description provided.