Problem
The documentation currently overstates or misstates some implementation details:
- README says "R-Tree Spatial Indexing" but the code uses K-D trees.
- README suggests A*/Dijkstra routing, but current public route calls pass a zero heuristic, which behaves like Dijkstra.
- README promises "Full road-following geometries", while the main
route APIs still snap to nearest nodes.
Why this matters
This creates user confusion and makes correctness/performance expectations unclear.
Scope
- Update README feature list and API examples.
- Clarify the difference between node-snapped and edge-snapped routing if both remain public.
- Document current complexity honestly until the implementation changes.
Acceptance criteria
- README terminology matches the implementation.
- API docs explicitly describe which public methods are node-snapped vs edge-snapped.
- Performance claims are accurate and non-ambiguous.
Problem
The documentation currently overstates or misstates some implementation details:
routeAPIs still snap to nearest nodes.Why this matters
This creates user confusion and makes correctness/performance expectations unclear.
Scope
Acceptance criteria