상태에 반응하는 모션

Primary role: motion behavior catalog.

Repository Boundary

These recipes describe product behavior and failure cases. They compose with Layout instead of supplying reusable spatial CSS or visual defaults. Start with a Motion Brief; use Vocabulary for the visual effect's name.

Reusable Method

Choose the task below, assign semantic and presentation owners, implement the immediate state change first, and then test any visual interpolation against that baseline. The traces below are expected behavior, not evidence of a tested implementation.

Action Feedback

Use for a command whose acknowledgement and completion are different events.

ready → pending → success | error; a retry is a new request. Press feedback may acknowledge input immediately, but a success treatment waits for actual confirmation. Decide explicitly whether repeated activation is ignored, queued, or replaces the request. Keep essential status available without motion. Test delayed completion, duplicate input, failure, and a response arriving after navigation.

Disclosure

Use for optional content controlled by a nearby trigger. A native details/summary baseline can be enough:

<details>
  <summary>Delivery restrictions</summary>
  <p>Some destinations need an additional delivery day.</p>
</details>

If a custom disclosure is necessary, keep the trigger's expanded state consistent with content availability. The APG disclosure contract describes button activation and state semantics. When closing while focus is inside, move focus to an available logical target before hiding it. Test repeated toggle, long content, and removal of the trigger. Avoid fixed maximum heights that clip real content; Layout owns intrinsic sizing.

Modal Transition

Use when a temporary task must block the surrounding interface. Assign one owner to opening, focus placement, blocking, cancellation, and focus return. An entrance effect may decorate this change; semantic availability must not wait for an arbitrary timeout. If exit animation is retained, define whether the dialog remains modal until cleanup and how reopening supersedes a pending close.

The APG modal dialog guidance supplies the web interaction baseline. Test close during entry, reopen during exit, nested overlays, trigger removal, and reduced motion. After final close there must be no surviving blocker or focus trap.

Reorder And Filter

Use when the same identifiable items move or are removed. Commit the intended logical order, retain stable item keys, and interpolate presentation only where the source/destination mapping is valid. Preserve focused items; if one disappears, choose a documented neighboring item or controlling filter. Do not reorder DOM solely to produce a pleasing stagger.

Trace: filter A → filter B → late A data → B remains current. Test empty results, long item labels, list updates during movement, keyboard navigation, and a focused item's removal. An immediate redraw is the fallback when continuity cannot be established.

Progress And Completion

Use when an ongoing operation needs visible status. Show determinate progress only when backed by a meaningful measurement; otherwise use a pending state. Completion, failure, and cancellation come from the operation, independently of the progress decoration.

Trace: pending → cancellation requested → confirmed cancelled or completed; do not promise cancellation succeeded before the operation confirms it. Essential information persists beyond a transient flourish. Use urgent alerts sparingly; APG alert guidance distinguishes announcements from focus-taking dialogs.

Drag And Settle

Use when direct manipulation previews a change before commitment. Separate the preview value from the committed value. Release on a valid target commits; cancellation restores or reconciles with the current model. A new gesture retargets from the presented position where continuity is intended. Provide a task-equivalent non-drag operation.

The non-drag operation has two separate audiences. WCAG 2.2 Success Criterion 2.5.7 Dragging Movements (Level AA) asks that functionality operated by dragging also be achievable with a single pointer without dragging: a click or tap, not a held path. Its Understanding document states that keyboard equivalence does not meet this criterion by itself unless those controls can also be clicked or tapped, and that the keyboard and single-pointer requirements are evaluated independently. A sortable list with arrow-key reordering therefore still needs pointer-operable move controls, and move buttons that only respond to a pointer still need keyboard operation.

Two exceptions bound the criterion. Dragging may remain the only pointer method when it is essential to the task, and the criterion does not cover dragging that the user agent determines and the author has not modified, such as native scrolling. The alternative cannot itself be a path-based gesture such as a swipe or flick; the Understanding document's own examples are tapping a slider track, move controls beside a sortable item, and a text input for a precise value. It does cover content that suppresses native scrolling and interprets the drag itself, which includes a custom carousel, a canvas pan, or the art-surface swipe in full-viewport scene navigation. Record an essential claim with its reason rather than assuming it.

Failure cases: the alternative exists only in a keyboard handler; the alternative appears only on hover, so touch cannot reach it; the alternative commits a different value or skips the validation that release performs; the alternative moves the item but leaves focus or the announced position on the old location.

Test release inside/outside, cancellation, pointer loss, a changing destination, and repeated reversal. Run the same task three ways (drag, single-pointer activation without movement, keyboard only) and compare the committed value and the final focus. Focus and the selected item must remain traceable while visual position changes. The engine or browser adapter owns gesture events; the product owns acceptance of the new value.

Scroll-Driven Story

Use when a visitor explores a product through ordered visual chapters: scroll a phone into view, enlarge the device, change its screen, then release the scene into the next section. The story is presentation; scrolling must not commit a payment or other application action.

The Scroll Story Lab links runnable Flow and headphones examples, a native CSS scrub, source files, and observed verification. It implements scrub, pinning, scene crossfades, and a small image sequence. Parallax and video seeking remain contracts rather than claimed executed examples.

Scene Composition Contract

Layout owns document scroll, the containing block, sticky offset, available stage height, and track distance. Motion owns progress, scene ranges, continuity, and teardown. The consuming product owns imagery, typography, colors, and visual values. First compose stack, content-limiter, and overlay-stack responsibilities with local sticky positioning; a new Layout pattern requires a distinct demonstrated spatial failure.

Record start, distance, scene_ranges, resize_policy, direct_entry, static_path, and media_budget alongside the Motion Brief. For positive distance, clamp (scroll_position - start) / distance to [0, 1]; a nonpositive distance selects the static path. Derive the entire scene from current progress, not from which callbacks have already fired. This makes reversal and skipped segments well defined.

The worked three-scene example centers chapters at progress 0, 0.5, and 1. Copy crossfades between centers while decorative scale and position interpolate. These values are local choices, not universal timing or easing defaults. No wheel interception or artificial scroll smoothing is needed for the example.

Implementation And Fallback

Choose the smallest rendering mechanism that meets the contract. CSS scroll/view timelines suit declarative interpolation; a scheduled JavaScript renderer can coordinate semantic availability and decoded media. Feature detection only chooses an implementation candidate, not a supported-device claim. The lab demonstrates CSS continuous rotation separately from its JavaScript three-scene controller; it does not claim visual parity between them.

Keep all chapters in ordinary semantic HTML before enhancement. In animated mode, only the presented chapter is available to interaction and accessibility APIs; a visible reading-mode control exposes every chapter in order. Provide a skip link outside the changing panels. Do not put essential actions only inside a fading chapter. Narrow or low windows, overflowing copy, no script, and reduced motion must retain a readable path. React to preference changes during use.

Cache geometry on layout changes, batch scroll updates, and avoid a perpetual idle animation loop. Recompute on resize, delayed fonts/images, and history restoration. Dispose observers, queued frames, and stale media callbacks on removal. A restored page must recompute from its restored position.

Media Contract

Record frame order, dimensions, poster, compression, source/license, decoded-memory estimate, cache limit, concurrent decode limit, timeout, and fallback. A successful older decode must never replace a newer requested frame. Retain a poster or last good frame while loading or after failure. Keep essential text outside the image. For video, additionally define duration readiness, seek granularity, pending-seek replacement, and seek failure; selecting currentTime is not proof of frame-accurate delivery.

The lab uses twelve locally drawn 480×480 SVG illustrations, one in-flight decode, three cached frames, and an eight-second example timeout. These low-detail illustrations verify the loading lifecycle, not photographic quality or production video performance.

Verification Trace

Exercise 0 → 0.5 → 1 → 0.5 → 0, direct entry at 0.5, and jumps across both scene boundaries. Compare presented state at the same progress in both directions. Also exercise resize mid-scene, text growth, image failure, a slow obsolete request, idle updates, reading-mode activation, preference changes, and teardown. Record browser and OS separately from emulated viewport/input. A screenshot proves one state; performance and perceptual continuity need separate evidence.

Technical sources rechecked on 2026-09-09: W3C Scroll-driven Animations draft, requestAnimationFrame, and image decode. These support the API distinctions and decode lifecycle; they do not establish local frame rate, comfort, or adoption.

Full-Viewport Scene Navigation

Task: explore a short sequence as screen transitions while the document stays at one position. This is discrete navigation driven by input, not a scroll-position timeline. Use the Scene Navigation example for a runnable contract.

Choose explicitly; similar appearance does not make their input, history, or recovery behavior equivalent. A short showcase may justify scene navigation. Articles, search results, forms, and long task flows should retain ordinary reading and scrolling unless a tested task requires otherwise.

Input And Navigation Contract

One controller owns the selected chapter. Commit that state immediately and let product CSS retarget presentation; never wait for an animation timer before accepting navigation. Keep the same decorative subject node across scenes when object continuity is intended.

• Declare the input surface. The example owns vertical wheel gestures only within its stage and touch swipes only over its decorative art; visible chapter and previous/next controls remain available.
• Aggregate small wheel deltas and accept at most one navigation per burst. Declare threshold, unit conversion, silence interval, reversal, and edge behavior as local choices. Test long momentum tails on actual target hardware before claiming trackpad support.
• Exempt editable controls, native overflowing copy, modifier gestures, and horizontal scrolling. Do not capture Tab, browser history shortcuts, or pinch zoom. Do not auto-advance from a copy pane when its scroll reaches the edge.
• Support direct chapter links, keyboard commands, Back/Forward restoration, and invalid-fragment fallback. Do not loop at the first or last chapter. Preserve unrelated URL state.
• Make only the selected panel interactive in scene mode. If focus was inside the outgoing panel, move it to the incoming heading; otherwise retain control focus. Announce the chapter separately from animation.
• Provide a persistent reading-mode control and an escape command. All chapters remain ordinary semantic HTML with no script. A short viewport selects reading mode; long copy remains a named native scroll region.
• Reduced motion may preserve discrete navigation while making transitions immediate. State this separately from the pinned example's static fallback, and react to preference changes during use.

The example's 60-pixel aggregate threshold, 180-millisecond silence interval, 55-pixel swipe threshold, and 540-pixel minimum height are product experiments. They are not gallery-wide defaults or measured usability recommendations.

Verification Trace

Record 0 → 1 → 2 → 1 → 0 with document position unchanged, a long wheel burst, fresh reverse gesture, rapid opposite commands, boundary commands, direct entry, Back/Forward, and Escape. Exercise long copy and text editing without changing chapters, cancelled/multi-touch gestures, no script, reduced motion, resize, and teardown. Desktop touch emulation does not establish iOS Safari, Android, physical trackpad, or assistive-technology behavior.

Locally authored interaction contract. The worked example's verification is bounded to its recorded source and runtime; no external homepage implementation is inferred from its appearance.

Opinionated Guidance

For frequent operations, compare with removing motion before increasing its complexity. When two recipes compose, one controller must arbitrate the shared state rather than letting competing completion callbacks decide it.

Platform-Specific Guidance

For the web, prefers-reduced-motion is a preference signal. Define the alternative per recipe and respond to preference changes during use. API support, native gesture physics, and rendering performance require the exact target runtime.

Unsupported Absolutes

These recipes establish no universal duration, spring configuration, frame rate, device coverage, or aesthetic result. A static snippet does not prove animated or assistive-technology behavior.

Verification Contract

For each implemented recipe record before, intermediate, interrupted, and final states; logical order and focus; reduced-motion behavior; and completion/error cleanup. Source inspection can check declared ownership. Rendered interaction and assistive-technology behavior require the consuming application. Review when an API changes or a trace fails.

The Interaction Lab provides a bounded web implementation of feedback, native disclosure, modal dismissal, stale search results, reward reveal, and range-input preview. Its verification notes record exercised states and preferences; they do not establish coverage for every possible implementation of these recipes.

Source, License, And Attribution

Locally authored recipes and HTML. The linked W3C APG pages and Media Queries Level 5 were rechecked on 2026-09-08 for the specific semantics and preference signal above. WCAG 2.2 (W3C Recommendation, 12 December 2024) and its Understanding page for Success Criterion 2.5.7 (updated 10 August 2026) were rechecked on 2026-09-19 for the criterion text, its two exceptions, and its relationship to keyboard operation; the Understanding page is informative, and the failure cases above are local and unexecuted. No upstream examples or timing prescriptions are copied.

IA Navigation

Parent: Motion. Next: Motion Review Workflow.