A focused audit of the four-node flight into the Busabase workspace sidebar, observed at 8.307 seconds. The concept is strong; the landing implementation breaks object identity at the exact moment the transformation should become legible.
The idea reads: external tools become trusted workspace nodes. The strange feeling is implementation-specific. A source chip and a destination row are visible in different places at the same time, while four easing segments repeatedly change the chip's velocity. Fix identity first, then shorten and simplify the wave.
Weighted for a polished product video: continuity and production finish first, speed and restraint second, experimentation only where it strengthens the one-time transformation.
| Lens | Verdict | One-line read |
|---|---|---|
| Production PolishJakub KrehelPrimary | Problem | The landing is a crossfade between two different objects, not a continuous transformation. |
| Restraint & SpeedEmil KowalskiSecondary | Concern | A 580ms flight is split across four eases and restarted every 240ms, making the wave feel fussy. |
| Experimentation & DelightJhey TompkinsSelective | Strong | The converging wave is worth keeping; it is the right place for expressive motion in this scene. |
The total 580ms can work for a cinematic transformation. The problem is its internal structure: four speed resets, a 101ms dual-visible landing, and a new node launched before the prior one settles.
At 8.307s the App source is about three quarters through its child fade, while the App row is already near its final opacity. Both remain readable because they occupy different coordinates.
Problem
14-trust/en/index.html:33214-trust/en/index.html:538WhatThe source node's children fade from at + .40 to at + .58. The destination row begins entering at 44% of that same handoff and is nearly opaque before the source is hidden. At 8.307s the old App chip and the new App row are visibly separate.
Why it mattersThe scene's claim is continuity: work from each tool becomes a Busabase node. Two simultaneous objects read as duplication, not transformation.
Recommended motionKeep one source icon alive through the entire flight. Reveal the row shell and copy during the last 120ms, but keep its icon hidden. At the exact landing coordinate, swap source icon to target icon over one or two frames. A FLIP/shared-element handoff is ideal; a deterministic overlay proxy also works.
14-trust/en/index.html:544-545 and 14-trust/en/assets/motion-system/index.js:55-60
WhatThe media query disables CSS transitions on later evidence elements only. The registered GSAP timeline still runs the large camera move and all four node flights.
Why it mattersThis is full-frame zoom and translation. The audit contract treats a missing reduced-motion alternative as critical, especially for a browser-previewed composition.
Recommended motionBuild a second seek-safe timeline under gsap.matchMedia('(prefers-reduced-motion: reduce)'): set the camera to the workspace framing, reveal the four rows with a short opacity-only sequence, and skip all large x/y/z travel.
14-trust/en/index.html:100 and 14-trust/en/index.html:537-547
14-trust/en/index.html:332-335The vocabulary is already polished; the identity contract is not. Finish the handoff as one continuous object and the same design will read as intentional.
Concern
14-trust/en/index.html:547WhatThe same node moves for 80ms with sine.inOut, 180ms with power2.in, 180ms with power2.inOut, then 140ms with power3.out. Scale and z also change at every boundary.
Why it mattersEvery boundary creates a new acceleration profile. Near the sidebar this reads as a shrink, tug, and settle rather than a single confident arrival.
Recommended motionUse one 420ms continuous transform with one strong in-out curve. Put the arc on a motion path or a nested y-channel, not four sequential tweens. Scale once from 1 to the exact destination icon scale, then perform the identity swap at rest.
14-trust/en/index.html:539-543
14-trust/en/index.html:547Keep the wave, remove the internal ceremony. One curve and one landing communicate more than four mini-beats packed into half a second.
Strong
No additional issue from this selective lens.
Do not remove the playful convergence. Make the transformation technically cleaner so the expressive idea is what viewers notice.
Identity first, then comfort, then speed. The source files for English and Chinese mirror this implementation and should change together.
| Issue | File | Fix |
|---|---|---|
| Source chip and destination row are simultaneously visible in different places | en/index.html:544 | Shared-element landing; swap icons only at the same coordinate |
| Reduced motion does not affect the GSAP flight and camera | en/index.html:100 | Alternate opacity-only GSAP timeline via matchMedia |
| Issue | File | Fix |
|---|---|---|
| One flight restarts velocity across four easing segments | en/index.html:539 | One 420ms curve or shared-layout tween |
| Enhancement | Where | Impact |
|---|---|---|
| Reveal destination shell and copy separately from its icon | en/index.html:332 | Makes the transformation legible without duplication |
| Shorten flight while preserving the wave | en/index.html:547 | At most two simultaneous transfers; cleaner next-scene handoff |
Jakub Krehel - Production Polish
The visible defect is a craft gap: two representations appear where viewers expect one continuous object. Jakub's shared-layout and subtle staging lens therefore drives the first fix; Emil simplifies the speed profile; Jhey protects the useful expressive concept.