Merge Logic#

The merge layer combines local observations with the server‑declared global state to produce a single Awareness value. This value reflects both the device’s immediate environment and the broader conditions published by the RTT ecosystem. The merge is intentionally simple, predictable, and transparent.

Inputs to the Merge#

The merge logic receives two independent inputs:

• Local State — derived from network clarity, device stability, and runtime behavior.
• Server State — declared at /.well-known/rtt-awareness and cached by the app.

Each input is reduced to a coarse stability classification before merging.

Merge Principles#

The merge follows several principles to ensure consistent behavior:

• Local and global signals are peers — neither overrides the other.
• Instability is meaningful — drift in either domain affects the final state.
• Stability requires agreement — both sources must be stable for the app to report Clear.
• The model is declarative — no inference or scoring is performed.

These principles keep the merge logic predictable and reviewer‑safe.

Merge Outcomes#

The merge produces one of four Awareness states:

• Clear — both local and server signals indicate stability.
• Local Drift — local instability with a stable global state.
• Global Drift — global instability with stable local conditions.
• Drift — both local and global signals indicate instability.

These states form the basis for the Awareness indicator shown in the UI.

Decision Rules#

The merge logic can be expressed as a simple decision table:

• Local stable + Server stable → Clear
• Local unstable + Server stable → Local Drift
• Local stable + Server unstable → Global Drift
• Local unstable + Server unstable → Drift

This structure ensures that the app’s Awareness state always reflects both immediate and global conditions.

Caching and Fallback Behavior#

The merge layer uses the most recent server state available:

• Cached values are used when offline.
• Cache expiration is conservative to avoid rapid state changes.
• Local signals continue to update even when server data is stale.

This approach maintains stability during degraded network conditions.