Measured semantic-resource and dataset experiments

Resource scope and canonical observations

The new analysis retains 2,408 evidence-linked assertions for 86 resources across 28 requested dimensions. Known legacy scope judgments map to an ordinal rubric; new unsupported dimensions remain null. Zero means not-targeted in the assessed scope, not proven absence of a concept. All scope scores remain PROJECT ASSESSMENT, with provisional confidence and no fabricated expert review.

Twenty-five observations have explicit requirement IDs. Exact best-set enumeration through size three yields size 1: 37.2% (eml); size 2: 52.7% (ebv, eml); size 3: 62.2% (ebv, eml, foodon). The objective counts repeated dimension requirements with at least meaningful assessed scope. It does not measure exact term expressibility, equivalence or ontology quality. Longer accumulation curves are greedy witnesses; they do not prove minimal sets beyond size three.

Exact SOSA term presence is measured separately in canonical_term_witnesses.json. Presence is a building-block witness, not an approved scientific mapping. Subdomain accumulation, marginal gain, resource-pair overlap, unknown dimensions, threshold failures and complete-case exploratory PCA are retained independently. No missing values are imputed as zeros.

Real graph structure

Local snapshot	Named classes	Longest named-class depth	Median shortest root depth	Restrictions
bfo	36	5	4.0	31
sosa	16	0	0.0	0
ssn	11	0	0	73

These are explicit local triples, not imported or inferred closures. SOSA/SSN modeling patterns and external superclass references are not summarized by local named-class depth. SSN's explicit SOSA import is verified; other inventory relationships retain their weaker declaration provenance. Identifier resolvability, I-ADOPT compatibility and semantic crosswalk correctness remain unmeasured.

Candidate dataset diagnostics

The 18 records include product/network descriptions, a standard comparator and an unselected climate counterpart. They are not 18 integrated datasets. Each source statement has primary evidence and scope; dimension and field-presence coding are separately labeled PROJECT ASSESSMENT. Undocumented fields remain unknown.

Semantic diversity is a documented dimension lower bound, not variable count. The ontology-potential diagnostic counts proposed resource relationships and is a PROJECT ASSESSMENT. Formal ontology references are distinguished from standard use. Eight separate review-prioritization rubrics and all 153 pair diagnostics are available. Ties have no scientific standing; mixed documentation scope confounds ranking. Pairwise added dimensions do not grant scientific compatibility.

Evidence inventory · Canonical results · Graph measurements · Dataset and pair diagnostics.