Physical environments — evidence and remaining work

Meaning update, 5 October 2026: all four environment recommendations are approved, alongside the separately accepted digital-system and icon/component choices. The evidence below retains its original bounded preparation scope. Exact placement, vocabulary and implementation remain open.

History and present scope

The original spreadsheet's four environmental entries and the August extraction request were inspected. The request asks for selective extraction and review, not blanket adoption. The B-05 card bundled those entries with digital identity. Its implication that a store design necessarily establishes an identity system is not warranted by the examples. This review narrows that claim while preserving the original card as history.

Current rulings already support photographs linked to depicted work, independently useful parts, multiple discovery paths and source statements distinct from interpretation. Those capabilities are not asked again. The July strategy proposal also records the distinction between source-documented work and analysis, and service-side reachability through shared meanings. Its superseded mirror-row implementation must not be revived from the earlier section. The present recommendations do not depend on approving that obsolete mechanism.

The Installations and Signs packet separately owns piece types and screen presentation context. That packet is now answered. These environment descriptions complement its accepted types without reclassifying every photograph or media file. No count of unresolved entries is inferred merely from missing intake answers.

Inspected examples

• Endpoint's John Lewis signage: primary text and its directory/wall-graphic image inspected in the preceding evidence unit. A later text fetch failed; this is retained dated evidence, not a fresh successful fetch. The source supports a signage contribution, not credit for the whole interior.
• Apple's 2016 Union Square announcement: primary text inspected in the browser. It describes several coordinated areas. Used as a documented retail-design example, not a current site measurement or image-based classification.
• Pentagram's Mercury booth: primary text and the actual page photograph inspected. Blue-and-white structures and screens are visible. No original screen video was identified or played.
• Endpoint's Aegis Media graphics: primary text and workplace image inspected in the preceding evidence unit. The source distinguishes the graphics commission from the interior architect. The image alone cannot establish every contributor or the reusable rules of an identity system.

The recommendations compare keeping the older four names, using only broad Environmental Design plus setting, and retaining precise descriptions while narrowing Workplace Branding. The last approach is recommended. These are proposed browsing meanings; their registry placement requires consumer analysis, not another duplicate taxonomy by default.

Selected technical evidence

The repository receipt environment-identity-preflight.json preserves five literal read-only SQL responses and seven hashed local excerpts. Its selected schema checks cover the craft registry, guideline sections, content and the content-to-discipline junction. They do not cover every constraint, index or consumer.

The dated registry result contains active Environmental Design and a separate physical-activation service description marked for credits. Their presence is evidence of existing storage, not approval of every extracted identity element. A selected DevDay record is an event photograph of wayfinding work. That label describes the photograph and does not establish the depicted work's identity or all its links. A lexical guideline query found Environmental Commitments; that is a false match for spatial design, not evidence that no environmental guidance exists elsewhere.

The inspected local post-processor resolves active discipline terms and writes analyzer-provenance links. Its replacement path deletes earlier analyzer links before separate insertion steps. This raises a failure-atomicity question for the comprehensive audit; no failure was reproduced here. The selected frontend query uses discipline links for filtering. Neither observation proves the deployed writer, all readers, tracker equivalence or correct attribution.

The relationships the implementation audit must test

A source case study may document a project, a designed space, several contributing organizations and many photographs or original assets. Those are different objects and assertions. The source page must not force them into one content type or give every contributor authorship of every component. A location is not a campaign; a design reused at several locations is not automatically several new creative works.

Trace collection and capture through extraction, candidate identity matching, evidence, classification, validation, writes, publication and retrieval. Check that a photograph can find the depicted design, and that the original file can find supported uses without inferring them from shared source membership. Review joins and permission checks at each transition. Search, saved tracker rules and cited answers must preserve the same meanings.

Test changing installations, redesigns, uncertain attribution, removed sources and human corrections. Historical captures and prior citations must retain what they actually support. Monitoring may discover another appearance without authorizing a new classification run or overwriting a human decision. Scheduling, cost limits and delivery need separate live verification.

The full audit still owes data types, required values, primary and foreign keys, uniqueness, relationship direction and multiplicity, junctions, indexes, merge/deletion behavior and representative query plans. It also owes deployed prompt and writer verification, correction preservation and end-to-end examples. These outstanding checks do not turn settled human meanings into new approval questions.

Prepared 5 October 2026.