One scientific state.
Many runtimes. One contract.
+A consolidated operational view of the qFoldIT organization: scientific backends, gameplay patterns, SOS/SKG/SEM/UAG/UWI/MCP layers, engine adapters, the desktop standalone runtime, assets, CI, and repository inventory. Metrics are separated into measured source-of-truth data and declared architecture.
+ + +Canonical architecture
Four scientific gameplay pillars
Protein Folding → Construction
Amino-acid geometry becomes an assembly problem. The scientific metric remains the source of truth; the construction layer visualizes and exposes the consequences of player decisions.
Safety → Stress Test
Candidate molecules move through measurable safety and efficacy endpoints. Failure is diagnostic and feeds the next optimization step rather than becoming a cosmetic game-over.
Sequencing → Resource Discovery
Sequence discovery is framed as exploration and collection while preserving the distinction between a game representation and an authoritative sequencing computation.
Procedural Biology → Growth
Formal rewriting rules become a visible growth system. Players author rules, observe emergent structure, and connect the generated geometry to biological or ecological quantities.
Scientific gameplay compile matrix
| Pattern | Scientific backend | Mechanic | STANDALONE | Gate principle |
|---|---|---|---|---|
| protein_folding_construction | protein_design_mcp | construction | SUCCESS | theme only after verified backend |
| quantum_ground_state_portal | qfoldit_quantum_lab | portal_exploration | SUCCESS | scientific state is canonical |
| hp_lattice_spatial_puzzle | qfoldit_quantum_lab | spatial_puzzle | SUCCESS | failure remains measurable |
| admet_safety_strategy | qfoldit_admet_lab | optimization | SUCCESS | safety endpoints remain scientific |
| l_systems_procedural_construction | qfoldit_plant_growth_lab | construction | SUCCESS | growth model remains authoritative |
Runtime adapters
| Target | Scene | Procedural geometry | Physics | Interaction | Scientific visualization | UAG |
|---|---|---|---|---|---|---|
| UEFN-TOOLBELT | supported | partial | supported | partial | partial | supported |
| UNITY-TOOLBELT | supported | supported | supported | supported | supported | supported |
| UNIGINE-TOOLBELT | supported | supported | partial joints | supported | supported | supported |
| QFOLDIT-STANDALONE / Web runtime | supported | supported | supported | supported | supported | supported |
STANDALONE release surface
Electron shell
The standalone catalog distinguishes bundled executables from verified external/skill references. This keeps offline claims precise and prevents accidental network assumptions.
UAG bridge
The adapter dispatches through its registered tool registry, preserving the same contract as the engine toolbelts.
Scientific visualization
Rendering remains downstream of scientific state. The runtime can make science visible without becoming an unverified scientific solver.
Release gates
Organization repository map
Core / science
qFold · quantum-folding · QuPepFold · quantum-protein-structure-prediction · Atomic-Calculations · BIOmics · CAMEO-qFoldIT · nanover · NASA_MRET
MCP / scientific interface
MCP-Server · qFold-MCP · Protein-MCP · Protein-Design-MCP · Atomic-MCP · science-skills · Scientific-Object-Schema · ersilia-qfoldit-bridge
World / gameplay
Virtual-Lab · Virtual-Lab-Simulation · Game-MCP · GAME-WS-SERVER · qfoldit.github.io · PufferFish
Engine / toolbelts
UEFN-TOOLBELT · UEFN-VERSE-MCP · uefn-mcp · UNITY-TOOLBELT · UNIGINE-TOOLBELT · UNREALVIS · WEB-TOOLBELT · mmos-sdk-js
Infrastructure / deployment
AWS-Deploy · .github · .github-private
Delivery sequence
Canonical schemas
Keep SOS, SKG, SEM, UAG, UWI, adapter manifests and gameplay schemas versioned and machine-validatable.
Science gates
Register each scientific backend with explicit status, provenance, network requirements and licensing. Only verified/connected backends unlock themed experiences.
Runtime parity
Run the same UAG vectors and gameplay compiler against UEFN, Unity, UNIGINE and STANDALONE so portability is measurable rather than inferred.
Asset provenance
Treat characters, screenshots, 3D models, video, fonts, logos and external references as first-class inventory with provenance and license metadata.
Release automation
Install renderer dependencies in CI, build the desktop runtime, run conformance and gameplay compilation, then publish signed artifacts and a machine-readable release manifest.
Live dashboard
Generate metrics from repository manifests and CI results rather than hard-coded numbers, preserving a durable audit trail for every dashboard value.