36 lines
2.5 KiB
Markdown
36 lines
2.5 KiB
Markdown
# Ansh Week 4 Final Presentation
|
||
|
||
## 1. What I Explored In Weeks 1–3
|
||
- Camera preset reliability and state switching across Vanilla, R3F, and Thob.
|
||
- Scene state and step-flow orchestration (visibility, transforms, camera framing).
|
||
- Parent-child-grandchild hierarchy behavior and transform inheritance.
|
||
- Duplication behavior (independent versus linked semantics).
|
||
- Builder reliability signals from console patterns, especially duplicate binding registrations and camera-related instability.
|
||
|
||
## 2. What thob Does Well For Scene / Storytelling Flows
|
||
- Fast visual scene setup for non-code users.
|
||
- Reliable hierarchy behavior for parent-child-grandchild transform inheritance when nesting is authored cleanly.
|
||
- Reliable discrete state switching when using prebuilt state meshes and visibility toggles (validated strongly in Week 3).
|
||
|
||
## 3. Where thob Is Weak / Fragile
|
||
- Multi-target interaction orchestration is fragile: one trigger to many updates is not consistently atomic.
|
||
- Perspective Camera rotation reliability is a critical gap (objects can disappear when rotating camera axes).
|
||
- Storytelling polish is limited by missing transition primitives for smooth camera/transform blending between states.
|
||
|
||
## 4. My Most Important Recommendation
|
||
- Build a first-class orchestration layer where one trigger applies one bundled, transactional state change across camera, visibility, and transforms.
|
||
- This should include deterministic execution, clear validation feedback, and support for rapid repeated interactions.
|
||
|
||
## 5. My Best Quick Win
|
||
- Fix duplicate binding registration lifecycle issues (GetBindingData, update-static-component-prop, resetPOI patterns).
|
||
- This is the fastest high-impact reliability improvement for Task 2 and Task 3 workflows.
|
||
|
||
## 6. My Deeper Architecture Concern
|
||
- Current advanced flows depend on workaround architecture (prebuilt states + visibility toggles) rather than a native step/state model.
|
||
- Without a first-class guided-flow schema and runtime orchestration semantics, complexity grows quickly as scenes become richer.
|
||
|
||
## 7. If I Had 2 More Weeks
|
||
- Run a focused camera reliability investigation and regression suite for Perspective Camera rotation behavior.
|
||
- Prototype a Guided Flow Step block with atomic multi-target updates and reset behavior.
|
||
- Add transition controls (duration/easing) and evaluate storytelling quality gains versus baseline instant switching.
|
||
- Stress-test orchestrated flows with higher object counts and deeper hierarchies to validate scalability. |