v0
Core Tech Validation (Headless-First)
Goal: Prove the hardest risks work in tests/desktop, no iPad hardware needed
Deliverable:
- CRDT sync converges across 4 simulated peers
- State persists and reloads correctly
- Session lifecycle (connect/disconnect/rejoin) works
- Synced cube entities maintain consistency
- All tests pass on macOS desktop
Scope: Desktop tests only, no iPad hardware required
v1
Advanced Engine Systems (iPad Hardware)
Goal: Prove complex systems work together on actual hardware
Scope:
- Agent Simulation: Foundation, Life Arc, Schedule, Behavior Trees, Performance, Networking (Phases 1-4, 11-12)
- Spatial Audio: Firewheel integration (Phase 1)
- Platform: Apple Pencil input validation
- RFC: Terrain Designer design doc
- Validation: 50 synced entities @ 60 FPS on iPad
Deliverable: iPad app with synced entities executing behaviors, test audio playback, pencil input. Debug visualization.
v2
Asset Pipeline & Rendering
Goal: Load real assets, custom rendering, basic animations
Scope:
- Rendering: Vendor Bevy renderer (Phases 1-4+)
- Assets: GLTF models, textures, basic skeletal animations
- Agent Simulation: Substrate, Institutions, Relationships, Conditions (Phases 5-8)
- Spatial Audio: Steam Audio, bus mixing (Phases 2-4)
- Terrain: Early terrain designer implementation
Deliverable: Real 3D models with animations, custom rendering pipeline, richer behaviors, 3D audio.
v3
Visual Polish & Advanced Features
Goal: Particles, advanced rendering, full simulation depth
Scope:
- Rendering: Particles, lighting, shadows, post-processing
- Agent Simulation: Tools, Production, full 200+ entities (Phases 9-10)
- Spatial Audio: Full feature set (Phases 5-8)
- Terrain: WFC generation
Deliverable: Polished visuals, full agent simulation, complete feature set.