A computational framework where physics emerges from information dynamics
Reality Engine simulates a universe where fundamental physics emerges from three coupled fields (Energy, Information, Mass) evolving on a Mobius manifold. No physics is programmed — no F=ma, no E=mc^2, no Schrodinger equation. Instead, 18 composable operators implement local rules, and physics appears as emergent behavior.
Based on Dawn Field Theory principles (PAC/SEC/RBF/MED).
Status: Pre-alpha research software. Not accepting contributions yet.
A 3-tier physics scorecard validates emergent behavior against DFT predictions:
| Coupling | Target | Measured | Error | Grade |
|---|---|---|---|---|
| f_local | gamma_EM (0.5772) | 0.643 | 11.4% | C |
| gamma_local | 1/phi (0.618) | 0.554 | 10.4% | C |
| alpha_local | ln(2) (0.693) | 0.675 | 2.6% | A |
| G_local | 1/phi^2 (0.382) | 0.343 | 10.3% | C |
| lambda_local | 1-ln(2) (0.307) | 0.325 | 5.8% | B |
| Metric | Status |
|---|---|
| phi^2 mass spacing | 10.8% error (C) |
| PAC conservation | Machine precision (A) |
| Spin half-integer | 11.1% error (C) |
| Entropy reduction | Arrow of time emerges (A) |
Fine structure (1/137), Koide ratio (2/3), mu/e mass ratio (206.8) — not yet in range.
Overall: 9/13 passing, GPA C+
git clone https://github.com/dawnfield-institute/reality-engine.git
cd reality-engine
pip install -r requirements.txt
# Watch fields evolve
python examples/field_visualizer.py
# Run physics scorecard (13 metrics, ~3 min on GPU)
python scripts/physics_scorecard.py
# Run tests
pytest tests/v3/ -vThe v3 architecture (2026-03-14) is a ground-up rebuild around composable operators:
src/v3/
├── engine/ Engine + FieldState + EventBus + Config
├── operators/ 18 physics operators (Pipeline chains them sequentially)
├── analyzers/ 6 measurement modules (Conservation, Gravity, Atom, Star, Quantum, Galaxy)
├── emergence/ 3 structure detectors (Structure, Particle, Herniation)
├── substrate/ MobiusManifold + Constants + Projections
└── dashboard/ FastAPI + WebSocket + Plotly.js (real-time visualization)
Each tick, the Pipeline chains operators: state_new = op(state, config, bus).
| # | Operator | Role |
|---|---|---|
| 1 | RBF | Recursive Balance Field — computes dE/dt |
| 2 | QBE | Quantum Balance — dI/dt = -dE/dt (PAC conservation) |
| 3 | Actualization | MAR-gated integration, pi/2 harmonic modulation |
| 4 | Memory | Mass generation (bulk + gradient seeding), quantum pressure, diffusion |
| 5 | PhiCascade | Fibonacci two-step memory for phi-spaced mass levels |
| 6 | Gravity | Spectral Poisson solver with cascade-depth tiling filter |
| 7 | SpinStatistics | Emergent Pauli exclusion from information cost |
| 8 | ChargeDynamics | EM-like forces from charge field |
| 9 | Fusion | Nuclear fusion in dense, hot regions |
| 10 | Confluence | Mobius antiperiodic projection |
| 11 | Temperature | Local T from disequilibrium gradients |
| 12 | ThermalNoise | Langevin stochastic forcing |
| 13 | Normalization | Soft-clamp fields, Landauer reinjection |
| 14 | SECTracking | Read-only entropy + SEC energy tracking |
| 15 | Adaptive | Self-tuning damping and timestep |
| 16 | TimeEmergence | Emergent time from disequilibrium |
Operators are composable — add, remove, or reorder any operator.
Gravity (the most developed operator):
- Spectral Poisson solver with amplitude coupling
- Cascade-depth tiling filter (DFT exp_36): local gravity strong, global gravity suppressed
- Entropy-coherence modulation: xi_s = I^2/E^2 determines coupling strength
- Result: web-like mass filaments instead of runaway clumping
Mass generation:
- Bulk: gamma_local * (E-I)^2 — mass where disequilibrium is large
- Boundary: gamma_local * |grad(E-I)|^2 / (1+M) — mass nucleates at structure edges
- gamma_local = (E-I)^2 / (E^2 + I^2) — emergent per cell, not hardcoded
Conservation: E + I + M = const at machine precision (< 1e-12 error).
reality-engine/
├── src/v3/ # ACTIVE — v3 composable pipeline
├── scripts/ # Diagnostics (physics_scorecard.py, diagnose_gravity.py, ...)
├── spikes/ # Research experiments
│ ├── coupling_drift/ # Gravity/memory optimization (6 spikes, 2026-03-16)
│ ├── atomic_emergence/ # Atom classification
│ ├── big_bang/ # Big bang evolution
│ ├── thermal_validation/ # Heat-information coupling
│ ├── law_discovery/ # Automated physics discovery
│ └── universe_evolution/ # Long-term structure formation
├── tests/v3/ # 138 tests
├── examples/ # field_visualizer.py
├── docs/ # Theory and guides
│ └── legacy/ # Archived v1/v2 documentation
├── proof_of_concepts/ # v2 POCs (001-007, reference)
└── [legacy dirs] # core/, dynamics/, conservation/, etc. — v1/v2 reference
| Repo | Role |
|---|---|
| dawn-field-theory | Theoretical foundation (51 experiments) |
| fracton | Infodynamics SDK (PAC/Mobius primitives) |
| dawn-models | AI architectures using DFT principles |
- Emergence over encoding — Physics must emerge from field dynamics, never be hardcoded
- PAC conservation — E + I + M = const at machine precision (< 1e-12)
- Mobius topology — Anti-periodic boundaries f(x+pi) = -f(x) on a Mobius band
- DFT constants — Xi = gamma_EM + ln(phi) = 1.05843 as global frame attractor
@software{reality_engine,
title = {Reality Engine: A Computational Framework for Emergent Physics},
author = {Groom, Peter Lorne},
year = {2025},
version = {3.0.0-alpha},
license = {AGPL-3.0},
url = {https://github.com/dawnfield-institute/reality-engine},
note = {Based on Dawn Field Theory principles}
}AGPL-3.0. See LICENSE for details.
Reality emerges. Physics discovers itself.
Last updated: March 16, 2026