docs: update week2 task4 docs
This commit is contained in:
parent
ee8e2c3342
commit
29f0625b8c
@ -6,6 +6,7 @@
|
||||
- Motion design layouts where one parent motion drives a cluster of elements.
|
||||
- Does Thob feel strong enough for this use case?
|
||||
- Partial. The mesh-inside-mesh pattern works for parent-child inheritance, but Perspective Camera rotation seems broken because objects disappear no matter which rotation axis is used.
|
||||
- For step-selection interactions, button-driven setup was not reliable end-to-end; radio-button switching between preconfigured states is a more dependable fallback.
|
||||
- What would improve the experience?
|
||||
- Fix Perspective Camera component rotation so scene objects do not disappear during camera rotation.
|
||||
- Add reusable presets for common parent-child motion examples.
|
||||
|
||||
@ -34,6 +34,7 @@ Build a step-based scene where a parent group is translated and rotated, while c
|
||||
- We can make things like this by following the simple mesh-inside-mesh design.
|
||||
- Rotate the parent object, and children inherit the rotation and move with the parent while maintaining offset distance.
|
||||
- This matches the same parent-child transform behavior used in vanilla and R3F.
|
||||
- For guided step switching, my first intuition was to use button-driven setup, but that interaction flow was not reliable enough to trust end-to-end.
|
||||
- Perspective Camera component rotation feature appears broken right now: no matter how much camera rotation is changed on any axis, objects disappear.
|
||||
-Builder steps:
|
||||
- Create a parent mesh/group container.
|
||||
@ -74,6 +75,7 @@ Build a step-based scene where a parent group is translated and rotated, while c
|
||||
- Keep parent/child group rotation for the main motion behavior.
|
||||
- Avoid rotating the Perspective Camera component for now.
|
||||
- Use camera position changes and look target adjustments instead of direct camera rotation.
|
||||
- For step selection UI, prefer radio-button switching between preconfigured states while camera rotation remains unstable.
|
||||
|
||||
## Suggested Improvement
|
||||
-What should improve in Thob?
|
||||
|
||||
@ -20,7 +20,7 @@
|
||||
5. Validate whether grouped motion plus camera framing can be repeated reliably.
|
||||
- **Complexity:** Easy for hierarchy behavior, Medium for complete camera-guided flow due runtime issues.
|
||||
- **Main limitation signals:** Runtime + Editor UX + Unknown investigation needed.
|
||||
- **Workaround status:** Partial workaround only (keep parent-driven motion, avoid direct camera rotation, prefer camera position/look target adjustments).
|
||||
- **Workaround status:** Partial workaround only (keep parent-driven motion, avoid direct camera rotation, prefer camera position/look target adjustments, and use radio-button switching for preconfigured step states when button bindings are unreliable).
|
||||
|
||||
## Console Warnings/Errors Seen (Deduplicated) and Probable Meaning
|
||||
|
||||
@ -63,5 +63,6 @@
|
||||
|
||||
- Task 4 core hierarchy behavior works in builder: parent rotation correctly drives child motion with preserved local offsets.
|
||||
- The major blocker is camera reliability: direct Perspective Camera rotation can make objects disappear, reducing feature parity confidence.
|
||||
- For guided step selection controls, radio-button switching between preconfigured states is currently more dependable than button-heavy interaction wiring.
|
||||
- Repeated binding-registration and controlled/uncontrolled warnings suggest editor/runtime instability that can amplify interaction and camera issues.
|
||||
- Product priority for this task should focus on camera rotation reliability first, then interaction-state stability for repeatable multi-step scene authoring.
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user