Native River Village world-context review

Both tested whole-scene placements fail. The smoke corrections and all native painting detail remain intact.

River Village v4: At the tested 154-world-unit width and [0.5,0.85] anchor, the tall v4 scene loses its upper house/smoke above the desktop Hex frame. Its painted river ends against canonical land instead of meeting the ocean, and the self-contained rock/soil island has a hard contact edge. Phone Hex retains the full upper house but still has the water/contact mismatch. Native artwork remains detailed; fix role, vertical framing and geography-aware placement before considering any artwork change.

River Village v5: At the tested 154-world-unit width and [0.5,0.85] anchor, v5 stays within both Hex frames, but its enclosed boat basin, waterfall, bridges and rock/soil island do not match the canonical shoreline. The painted river ends on the beach and painted land covers canonical water. Region makes the dense scene visibly distinct from the approved village-lane reference. Resolve canonical water/contact and subject scale/style in context while preserving the native painting.

The same canonical existing village is selected in all18 captures. Actual place-paint witnesses and Hex source URLs are recorded;reload retains selection and source. The control uses the approved village-lane artwork. The experiment supplies a single village candidate to the unchanged selector,so other existing villages in Region also receive that source;this intentional isolation is not a production distribution decision.

The first attempt exposed a manifest URL mismatch:asset approval normalizes /tile-library/,while canonical places require an exact source URL match. A browser-only /assets/__experimental-context-only/ alias supplies each exact PNG in the successful second experiment. No alias exists in production. The passive painter wrapper records calls without changing selection or draw arguments.

UNAPPROVED LOCAL EXPERIMENT. Synthetic browser-only manifest entries are test inputs, not source approvals or live integration. Existing deployed selector/painter and ordinary navigation run in a fresh canonical QA world. Only feature=existing village sources are substituted for the candidate scenarios. Geography, simulation and production files/world remain unchanged. No performance acceptance or lifecycle acceptance. No CPU throttling,constrained-network,memory-budget,construction-state or universal-device acceptance is claimed. The source native canvases are unchanged;normal screen projection is part of rendering.

Next:resolve the actual village role and compatible water/ground placement,then test vertical framing without reducing native artwork. A panel/portrait role or isolated building extraction is a proposal requiring its own implementation and review,not an approved disposition.

Illustration reference: Approved reference (native-alias-v2/control-1440-hex-initial.png)

Illustration reference: v4 experimental placement (native-alias-v2/v4-1440-hex-initial.png)

Illustration reference: v5 experimental placement (native-alias-v2/v5-1440-hex-initial.png)

Validation:86/86 catalog/native/art-boundary checks pass. Both root and isolated wiki builds/checks pass. Local Art Atlas verifies all8,347 metadata/checksum/inventory records,44 update hashes,two unchanged original hashes,production runtime hashes andfour candidate displays withfour expanded-finding captures,without browser errors. Only the two candidate metadata records change;all8,345 other records,all approval objects and all native-art identities remain exact.