Building-specific facade canvas — draw & place images on a building face (iPad)

[nigel-hall-codes]

Captured via /sdlc capture regarding #266 (SF Building Template Pipeline).

What: A new iPad authoring capability to select a single face of a specific building and treat it as a 2D canvas — freehand draw on it and place images (decals/stickers/photos) onto that facade, like a 2D paint/layout surface.

Scope boundary (explicit): This is building-specific only — it applies to individual, hand-chosen buildings keyed to a real OSM building (per #266's Building-Specific design category). It is NOT part of the reusable template/procedural system: no template parts, no material roles, no neighborhood palettes, no procedural placement rules. The output decorates one building, not a style.

Open design questions for the sysdesign pass:

• Face selection: how a building face is picked & identified (footprint edge → wall quad? UV space? keyed by OSM building ID + footprint hash like San Francisco Building Template Pipeline #266's building-specific overrides).
• Canvas model: coordinate space (meters vs normalized 0–1), layer ordering, freehand strokes vs placed image layers.
• Asset/storage format: how strokes + image placements serialize and round-trip through the Home PC Asset Server, and how Unity applies them to the generated building (decal projector? baked texture? quad overlay mesh?).
• Relationship to AI sign generator (San Francisco Building Template Pipeline #266) — does a placed image reuse that asset store, or is this a separate freeform layer?
• Real-world scale & calibration (reuse San Francisco Building Template Pipeline #266's two-point calibration?).

Captured via /sdlc capture. Next: /sdlc work #N (sysdesign → drafting).