# Savedata Parking Note

This project is no longer using original `savedat*.dat` analysis as an active work surface.

Reason:

- A save file only becomes meaningful after the game state that writes and reads it is understood.
- Current EXE work still needs stronger event VM, map transition, battle-entry, and runtime state evidence.
- Earlier savedata work produced many candidate offsets and route-proof blockers, but those did not resolve the core EXE execution questions and added noisy review surfaces.

Keep as low-priority clues only:

- Original save files appear to be 1274 bytes in the sampled corpus.
- Some editor-backed fields were tentatively identified: money, item counts, character stats, and scene selector bytes.
- The previous `selector 2:0` route-proof line should not be used as an active completion blocker.
- Public or synthetic savedata samples are diagnostic only. They are not proof of original route execution.

Do not restart savedata analysis until:

1. The original event VM route execution path is understood well enough to identify state mutation sites.
2. The browser/runtime prototype has a stable game-state model for maps, party, inventory, battle, and story flags.
3. A specific question exists, such as "which save offset stores this already-understood runtime field?"

When restarted, begin from a narrow before/after capture plan tied to one known runtime variable. Do not rebuild broad route-proof or selected-root investigations from savedata first.
