Update TECHNICAL_NOTES.md

TECHNICAL_NOTES.md

@@ -1,52 +1,67 @@
-# RFT-Ω Harmonic Validation Interface  
+# RFT-Ω Harmonic Validation Interface — v3
 
-Public endpoints for verifying **Rendered Frame Theory (RFT)** harmonic-stability metrics *(QΩ and ζ_sync)* under synthetic noise.  
-This Space provides an open, reproducible test harness for anticipatory-stability and drift-recovery mechanisms within the **RFT-Ω Kernel**.  
+Interactive demonstrator for **Rendered Frame Theory (RFT)** harmonic stability under controlled synthetic noise.  
+This Space provides a **reproducible test harness** for anticipatory stability (QΩ) and synchronization coherence (ζ_sync) with:
+- **Domain profiles** (AI/Neural, SpaceX/Aerospace, Energy/RHES, Extreme)
+- **Adjustable noise slider (σ)** to probe robustness
+- **Adaptive baselines** (light “memory”) and **range validation** (no false 1.0 spikes)
 
 ---
 
-### 🛰️ Live Endpoints
-- **UI:**  [https://rft-omega-api.hf.space](https://rft-omega-api.hf.space)  
-- **Heartbeat:**  [https://rft-omega-api.hf.space/api/ping](https://rft-omega-api.hf.space/api/ping)  
-- **Metrics:**  [https://rft-omega-api.hf.space/api/metrics](https://rft-omega-api.hf.space/api/metrics)  
+## How to Use
+1) Open the Space → **https://rftsystems-rft-omega-api.hf.space**  
+2) Select a **System Profile** (AI/Neural, SpaceX/Aerospace, Energy/RHES, Extreme Perturbation).  
+3) Adjust **Synthetic Noise Level (σ)** with the slider (0.00–0.30).  
+4) Click **Submit** → JSON output appears with live values for QΩ, ζ_sync, status.
 
----
-
-### 🧩 Expected Output
-Each JSON call returns harmonic-coherence values that fluctuate slightly to simulate stochastic environmental drift.  
-
-**Example responses**
+**Example output**
 ```json
-/api/ping →
-{"ok": true, "message": "RFT-Ω API online and stable"}
-
- /api/metrics →
 {
-  "QΩ": 0.845,
-  "ζ_sync": 0.793,
-  "status": "nominal"
+  "System": "AI / Neural",
+  "Noise Scale": 0.050,
+  "QΩ": 0.922,
+  "ζ_sync": 0.798,
+  "status": "perturbed"
 }
 ```
 
-**Typical ranges**
+---
+
+## What to Expect
+
+**Typical stable ranges (nominal conditions)**
 | Metric | Range | Meaning |
-|---------|--------|----------|
-| **QΩ** | 0.82 – 0.89 | Harmonic-stability factor |
-| **ζ_sync** | 0.75 – 0.88 | Synchronisation coherence |
-| **status** | `"nominal"` | System in equilibrium |
+|---|---|---|
+| **QΩ** | 0.82–0.89 | Harmonic stability factor (amplitude) |
+| **ζ_sync** | 0.75–0.88 | Synchronization coherence (phase) |
+
+**Status classification (qualitative)**
+- `nominal` — low variance; coherent equilibrium  
+- `perturbed` — moderate variance; coherent but stressed  
+- `critical` — high variance; edge-of-failure regime
 
-Repeated requests should show < 0.05 variance, demonstrating **stable harmonic response** under noise.
+**Noise guidance by profile (rough starting points)**
+- **AI / Neural:** σ ≈ 0.01–0.10 (training drift / GPU jitter)  
+- **SpaceX / Aerospace:** σ ≈ 0.03–0.12 (vibration / telemetry lag)  
+- **Energy / RHES:** σ ≈ 0.02–0.10 (grid oscillations / load steps)  
+- **Extreme Perturbation:** σ up to 0.30 (stress testing / failure modes)
+
+**Notes**
+- The kernel applies **domain-specific weighting** (QΩ vs ζ_sync importance) and a **soft adaptive baseline** so repeated runs can show mild learning/self-stabilization.  
+- Outputs are **clamped to [0, 0.99]** to avoid saturation artifacts and to reflect realistic bounded metrics.  
+- Repeated runs at fixed σ typically show **< 0.05 variance** in stable regimes.
 
 ---
 
-### 🧠 Validation Purpose
-External research or engineering teams (e.g., xAI, Starship Systems, RobustBench, GLUE) can use these endpoints to :
-- benchmark RFT harmonic stability under controlled perturbations  
-- examine QΩ / ζ_sync variance as a predictive-drift signal  
-- reproduce RFT’s low-variance behaviour in noisy inference environments  
+## Validation Purpose
+- **Benchmark harmonic resilience** under controlled perturbations (σ sweeps).  
+- **Study predictive drift signals**: observe how QΩ and ζ_sync diverge/converge as σ increases.  
+- **Profile-specific tuning**: compare AI vs Aerospace vs Energy domains with the same σ to see weighting effects.
+
+For collaboration (e.g., xAI/RobustBench/GLUE-style testing), this interface can be extended with dataset hooks and logging while keeping internal parameters sealed.
 
 ---
 
-### ⚖️ Rights & Contact
+## Rights & Contact
 All Rights Reserved under **RFT-IPURL v1.0** and the **Berne Convention (UK Copyright Law)**.  
-**Author / Contact:** Liam Grinstead  —  liamgrinstead2@gmail.com
+**Author / Contact:** Liam Grinstead — liamgrinstead2@gmail.com