Inferal Ontology Discovery Vocabulary

The Inferal Ontology Discovery Vocabulary provides a controlled SKOS subject scheme for finding semantic artifacts by the problem or subject area that motivates their use. Catalog publishers classify modules, profiles, and mappings with stable topic IRIs; discovery clients can then offer shared facets, query expansion, and understandable match explanations without embedding a private taxonomy in application code.

This is an unofficial 0.1 Editor's Draft maintained in the Inferal ontology repository.

Feedback should be sent to contact@inferal.com.

This specification defines conformance for a discovery vocabulary, a classified catalog entry, and a discovery client.

A discovery vocabulary conforms when every published topic is a named odisc:DiscoveryTopic and SKOS concept, belongs to the discovery scheme, has one English preferred label and definition, and forms a shallow acyclic hierarchy satisfying the SHACL profile [[SHACL]].

A classified catalog entry conforms when each dcterms:subject used for Inferal discovery names a concept from this scheme. An entry MAY have more than one subject. Classification does not assert that the artifact defines every term associated with that subject.

A discovery client implementation MUST preserve artifact kind, publication status, and provenance as facets separate from discovery subject. A client MAY expand a narrow topic through skos:broader, and MAY use preferred labels, alternate labels, and definitions for query matching. It MUST NOT present a similarity score as semantic entailment or evidence of ontology conformance.

Purpose and audience

An exhaustive ontology index assumes that a visitor already knows the vocabulary name, prefix, or publisher they need. Discovery begins earlier: a visitor may know that they need to trace personal data, represent version migration, or record decision evidence without knowing which artifact carries those semantics.

The vocabulary is intended for catalog maintainers, ontology authors, website and desktop-client implementers, and adopters exploring the Inferal ontology library. It provides one reviewed classification source for both public discovery and expert catalog tools.

Scope and non-goals

Discovery topics describe a reason for seeking an artifact. They are intentionally broader than ontology terms and narrower than arbitrary search keywords. The scheme owns labels, synonyms, definitions, and a shallow browsing hierarchy.

The scheme does not classify artifact type, maturity, authority, publication access, license, or implementation quality. It is not a replacement for the domain semantics inside a classified ontology. Absence of a subject does not prove that an artifact is unrelated to that subject under open-world RDF semantics.

Competency questions

Which semantic artifacts address a visitor's stated problem or subject?
Which broad discovery area contains a narrower topic?
Which controlled label or synonym explains why an artifact matched?
Can independent clients derive the same facets from published RDF?

Namespaces and reuse

Prefix	IRI	Role
`odisc`	`https://ontology.inferal.com/modules/ontology-discovery/`	Discovery scheme and topics
`skos`	`http://www.w3.org/2004/02/skos/core#`	Concept scheme, labeling, and hierarchy
`dcterms`	`http://purl.org/dc/terms/`	Catalog subject assignment

The module reuses SKOS [[SKOS-REFERENCE]] because its concepts are indexing subjects with preferred labels, synonyms, and broader-topic relationships. It reuses dcterms:subject [[DCMI-TERMS]] for classification rather than defining a discovery-only assignment property.

Conceptual model

odisc:DiscoveryTopic identifies an extensible indexing concept. Every discovery topic belongs to the odisc:DiscoveryTopics scheme. A top concept is a homepage-level browsing area. A narrow topic has exactly one direct broader topic in version 0.1.

The hierarchy is not a class hierarchy. Following skos:broader supports query expansion and grouping, but it does not type catalog entries or domain data as instances of the broad concept. Topics may overlap, and one artifact may be classified under several branches.

Discovery topic term reference

Data governance

odisc:DataGovernance groups artifacts concerned with governing how data is understood and used.

Topic	Discovery boundary
`odisc:DataClassification`	Categories, sensitivity, and classification evidence.
`odisc:ProcessingLineage`	Processing activities, inputs, outputs, policy, and mitigation.
`odisc:AccessControl`	Authorization, credentials, publication access, and capabilities.

Artifact evolution

odisc:ArtifactEvolution groups artifacts concerned with change across a governed lifecycle.

Topic	Discovery boundary
`odisc:Versioning`	Version identifiers, ranges, ancestry, and compatibility.
`odisc:SemanticChange`	Semantic differences, RDF deltas, reviews, and change sets.
`odisc:Migration`	Verified transformation and retraction between model versions.

Knowledge claims

odisc:KnowledgeClaims groups artifacts concerned with assertions and their interpretation.

Topic	Discovery boundary
`odisc:Evidence`	Claims, arguments, warrants, provenance, and defeaters.
`odisc:Confidence`	Uncertainty under an explicit interpretation scale.
`odisc:DecisionRationale`	Deliberation, options, consequences, and rationale.
`odisc:ScopedStatements`	Claims qualified by graph, source, time, or applicability.

Structural models

odisc:StructuralModels groups reusable mathematical and representational structures.

Topic	Discovery boundary
`odisc:Collections`	Sets and explicit membership occurrences.
`odisc:Graphs`	Graphs, hypergraphs, incidence, vertices, and edges.
`odisc:Ordering`	Precedence, comparison, adjacency, and contextual rank.
`odisc:Ranges`	Intervals, bounds, and endpoint inclusion.
`odisc:Localization`	Positions, reference frames, time, and location.

Business domains

odisc:BusinessDomains groups reusable models of organizations, markets, and industries.

Topic	Discovery boundary
`odisc:CompanyIntelligence`	Entities, ownership, control, funding, and company facts.
`odisc:CommercialIntelligence`	Offerings, customers, use cases, and revenue models.
`odisc:Insurance`	Insurance actors, authority, capacity, and risk transfer.
`odisc:Advertising`	Advertisements, campaigns, audiences, and platforms.

Software systems

odisc:SoftwareSystems groups artifacts concerned with operational software interfaces and representations.

Topic	Discovery boundary
`odisc:Authentication`	Credentials, API keys, OAuth, signatures, and webhooks.
`odisc:DataAccess`	Capabilities, integrations, traversal, and pagination.
`odisc:StructuredValues`	JSON values, objects, arrays, and RDF representation.
`odisc:TypeSystems`	Types, schemas, fields, and versioned API definitions.
`odisc:Conversations`	Threads, messages, transcripts, and tool-mediated dialogue.
`odisc:Materialization`	Transforms, identity policy, record mapping, and generated views.

Ontology engineering

odisc:OntologyEngineering groups artifacts concerned with semantic artifacts as governed products.

Topic	Discovery boundary
`odisc:OntologyDiscovery`	Finding semantic artifacts by problem, capability, and subject.
`odisc:OntologyPublication`	Publication state, audience, delivery, and stable access.

Classifying catalog entries

Publishers SHOULD assign the most specific topics that materially aid discovery. They SHOULD NOT add a broad topic when a narrow assignment already reaches it through skos:broader. Multiple narrow topics are appropriate when an artifact genuinely crosses concerns.

@prefix dcterms: <http://purl.org/dc/terms/> .
@prefix odisc: <https://ontology.inferal.com/modules/ontology-discovery/> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .

<https://example.com/processing-model/>
    a owl:Ontology ;
    dcterms:title "Example processing model" ;
    dcterms:subject
        odisc:DataClassification,
        odisc:ProcessingLineage .

Free-form source vocabulary belongs in search keywords. It should not be minted as a concept merely because one artifact uses the phrase. Alternate labels belong on a controlled topic only when they are stable synonyms for that topic.

Validation and entailment

The SHACL profile requires named topics, human-readable definitions, scheme membership, and exactly one hierarchy role: top concept or narrow concept. It rejects direct and indirect skos:broader cycles. Catalog-wide tests separately verify that published Inferal entries use only controlled discovery subjects.

@prefix odisc: <https://ontology.inferal.com/modules/ontology-discovery/> .
@prefix skos: <http://www.w3.org/2004/02/skos/core#> .

<https://example.com/broken-topic>
    a odisc:DiscoveryTopic, skos:Concept ;
    skos:inScheme odisc:DiscoveryTopics ;
    skos:broader <https://example.com/broken-topic> .

SKOS hierarchy supports navigation and query expansion. This module adds no OWL addendum because it needs no stronger class, identity, closure, or property entailments. SHACL validation does not prove that a catalog classification is substantively correct; maintainers review that editorial decision.

Extension and evolution

The scheme is open to reviewed topics. New topics MUST have a stable Inferal IRI, one concept boundary, human semantics, and a demonstrated discovery need across at least one catalog artifact. Existing concept IRIs MUST NOT be reassigned to a different meaning. Clients SHOULD tolerate additional topics and deeper hierarchies even though version 0.1 publishes a shallow hierarchy.