Native Twin Optimization Plan

This is a forward-looking plan for reducing frontend-to-backend geometry transfer during slicing and export.

Goal

Move toward a native scene twin in Rust so the frontend can send small state diffs instead of repeatedly staging large geometry buffers.

Key constraints

• Support editing in the frontend must stay smooth.
• Support fidelity must remain exact.
• The work should land after the stable beta path is complete.

Architecture direction

• Frontend owns live interaction and preview.
• Backend owns canonical slice-ready state.
• Model assets are loaded by identity rather than resent repeatedly.
• Support changes are transmitted as graph diffs with stable IDs and resolved coordinates.

Success criteria

• Less bulk geometry IPC.
• Better support-heavy export performance.
• Revision parity between frontend and twin before slicing/export.

Status

This is a roadmap note, not a current runtime contract.