Use of non-existent dct:partOf — should be dct:isPartOf

[tkuhn]

Problem

We use <http://purl.org/dc/terms/partOf> in several places, but that term does not exist. The DCMI Terms vocabulary (namespace http://purl.org/dc/terms/) defines:

• ✅ dct:isPartOf — "A related resource in which the described resource is physically or logically included." (X dct:isPartOf Y = X is part of Y) — this is the correct property for a part→whole / "belongs to" link.
• ✅ dct:hasPart — the inverse.
• ❌ dct:partOf — not defined. <http://purl.org/dc/terms/partOf> is undefined/non-dereferenceable.

(A bare part_of exists in other vocabularies — e.g. OBO/RO BFO_0000050 — but not in dc/terms/.) So any triple using dct:partOf is effectively malformed and won't be recognized by DC-aware tooling or reasoners. It should be dct:isPartOf.

Filing here as the most convenient place to track it, though it's not Nanodash-code-specific — there are no occurrences in the Nanodash repo source (the only DC part-relation in code is dct:isPartOf, used correctly in ExplorePage.java). The actual usages are in published nanopub artifacts:

• Query / view / template nanopubs (e.g. the part-About "Belongs-to" info query).
• Possibly the gen / spaces vocabulary and other repos.
• Possibly already-published data nanopubs that assert dct:partOf.

Proposed work

1. Enumerate all usages — sweep published query/view/template nanopubs (and vocab + related repos) for http://purl.org/dc/terms/partOf to establish the blast radius.
2. Correct the editable artifacts — supersede the affected query/template/view nanopubs to use dct:isPartOf, and update any consuming code that reads partOf.
3. Note the irreducible part — already-published immutable data nanopubs that assert dct:partOf cannot be edited, only superseded; existing ones remain wrong in the record. Decide whether/how to migrate.

Direction check

dct:isPartOf points part → whole (resource dct:isPartOf owningResource). Confirm each call site intends that direction (vs. dct:hasPart) when correcting.