← Documentation home

Canonical Markdown source · Oct 20, 2018

Linked Art 1.1 Implementation Evidence

linked-art/1.1-meeting-implementation-evidence.md · 74 lines · SHA-256 342d9277c7d7

Meeting: August 5, 2026

Prepared: July 28, 2026

Project: Meta Museum by Sun & Rain Works

This packet summarizes implementation experience relevant to three agenda

issues. It reports current behavior and questions for the group. It does not

present pending proposals as accepted Linked Art 1.1 practice.

Issue #362: limited API fields

Meta Museum can produce a server-defined brief record without accepting

arbitrary client field lists. The provisional brief contract:

  • retains canonical `id` and `type`;
  • retains bounded discovery fields when present;
  • sets `_complete: false`;
  • provides `_links.full.href` for deterministic full-record retrieval; and
  • does not mutate or replace the canonical full record.

Implementation: `src/contracts/linked-art-response-profile.ts`

Evidence: `tests/contracts/linked-art-response-profile.test.ts`

Question for the group: should 1.1 standardize profile names, discovery

mechanics, required fields, or only the completeness and full-link semantics?

Issue #637: confirmed non-equivalence

Meta Museum distinguishes a low-scoring or rejected candidate from a

curator-confirmed non-match. A confirmed decision requires:

  • the two compared entity identifiers;
  • curator identity and decision timestamp;
  • rationale and at least one cited evidence source; and
  • the reconciliation score when available.

The internal assertion emits no `distinct_from`, `notEquivalent`, or other

negative Linked Art property while the vocabulary and assignment pattern remain

unsettled.

Implementation: `src/contracts/reconciliation-negative.ts`

Evidence: `tests/contracts/reconciliation-negative.test.ts`

Question for the group: should the negative relationship be asserted directly

between entities, or qualified through an assignment that carries evidence and

responsibility?

Issue #780: provenance of unenumerated Sets

The fixture models a conceptual `Set` with `members_exemplified_by` pointing to

a representative `HumanMadeObject`. Acquisition provenance is attached only to

that exemplar. The Set has neither invented members nor physical ownership

history.

Fixture: `tests/fixtures/linked-art-1.1/unenumerated-set-provenance.json`

Evidence: `tests/quality/linked-art-1-1-agenda-fixtures.test.ts`

Question for the group: is exemplar-level provenance sufficient when the Set is

intentionally unenumerated, and should guidance explicitly prohibit copying

that provenance onto the conceptual Set?

Requested meeting outcomes

  1. Record the accepted terminology and constraints for each issue.
  1. Link the resolution or approved example.
  1. Keep unresolved terms gated after the meeting.
  1. Update the model reference before changing validators or public mappings.

Record outcomes in

`1.1-meeting-decision-ledger.json`(1.1-meeting-decision-ledger.json). Agenda

proposals remain separate from final decisions, and a resolved row requires a

Linked Art issue resolution link, target release, and explicit local action.