A correspondence to a vocabulary governed elsewhere is asserted, never imported

Answers Taxonomy

Where a locally governed term corresponds to a term in a vocabulary published by someone else, the correspondence is asserted along a declared relationship to an external term, which is a term of that other vocabulary named by the identifier it publishes and carrying the address that establishes it. The other vocabulary is never imported: none of its terms becomes a term of this one, no identifier of its is carried as a property of any term here, and nothing it revises moves a local meaning. A correspondence is admitted only where that vocabulary publishes a stable identifier for the term and its own published definition is the same thing said: where the correspondence would be a judgement, it is not asserted at all. The relationship says which kind of correspondence it is, makes an evidential claim and nothing stronger, and is a classification rather than a road, because a term governed elsewhere is a hub many terms would answer to.

What goes wrong without it

Adopting an industry vocabulary as the master buys comparability and hands over governance of the words: when the external scheme revises, local meanings move with it, and a term that carried a specific operational sense is reinterpreted by a committee somewhere else. Refusing it altogether buys the opposite problem, because the systems a practitioner actually runs are keyed on those schemes, and a vocabulary that cannot say what it corresponds to can only be read, never joined. A mapping is the arrangement that keeps both, and the knowledge-organization standards have held it as its own set of statements for exactly this reason rather than as a property of either side. Asserting it as a relationship to a term of the other vocabulary, rather than as a field on a term of this one, is what keeps the two vocabularies separable: the correspondence can be withdrawn without touching either, and a reader can see whose term is whose. The evidence bar is high on purpose, because a correspondence asserted loosely is a false public claim about someone else's words, and a practitioner who trusts it keys a system on it.

What proves it — 2

Each is a gate the build runs: a check inside a ring, a whole ring, or an eval beside them. A build stops at the first that refuses.

• registry integrityrelational claim
  Every relationship declares what kind of claim it makes, a hedging verb claims no more than it says, a typing is always constitutive because it says what a thing is, a mapping is always evidential and stops where it arrives because it records that two published definitions agree and reaches a vocabulary governed elsewhere, one verb carries one claim, and no claim class sits unused.
• ring:2the laws ring
  The registry keeps every law: identity, shape, admission, the relationships and their bounds, the journeys as machines, the measures as calculations, the addresses, and the kernel's names.

Derived from — 3

• W3CSKOS Simple Knowledge Organization System Reference
• National Information Standards Organization, the designated maintenance agency for ISO 25964, publishing the standard's scope, data model and schemasISO 25964-2 Information and documentation, Thesauri and interoperability with other vocabularies, Part 2: Interoperability with other vocabularies
• International Organization for StandardizationISO 25964-1:2011 Information and documentation, Thesauri and interoperability with other vocabularies, Part 1: Thesauri for information retrieval

Kinds it licenses

Relationships it licenses

exactly matches

Where this model stands

A correspondence is a relationship of its own kind, declared in the law beside the roads and the other classifications, carrying an evidential claim: it records that two published definitions say the same thing and asserts nothing about how either behaves.

External terms are a kind of their own, admitted as world records: each names the vocabulary and the term, carries the address that vocabulary publishes it at, and grounds on the entity the term is about, so a term about a kind of page grounds where a content type does.

Three correspondences are asserted, each to a class schema.org publishes and defines in the same words: the page of questions and answers, the dated editorial post, and the step-by-step guide, whose own definition here already named the class it is marked up as.

The correspondences answer a declared question, so nothing is held that no question needs: what a content type corresponds to in a vocabulary published by someone else.

What it does not claim

Only exact correspondence is declared. The hierarchical and associative kinds the interoperability standard distinguishes, where a local term is narrower or broader than a published one, wait for a reader who needs them: several would be assertable today and each would be a judgement, which is the bar this commitment sets against.

Only content types carry a correspondence. Surfaces, creative formats and audience segments are the ones a delivery or measurement system is keyed on, and each waits for a published identifier whose definition is the same thing said.

A correspondence carries no version of the vocabulary it points into. The address is stable and the definition behind it may be revised, so a reader cannot yet see which release was compared.

RUL-060