Ontology

What is an ontology?

An ontology is a formal description of the concepts that exist within a particular domain, the properties those concepts are expected to have, and the relationships that can hold between them. Where a single schema describes the expected shape of one kind of thing, an ontology describes a whole vocabulary of interrelated things, providing a shared frame of reference that both humans and machines can agree upon.

The term originates in philosophy, where “ontology” is the study of what exists. In information systems it has a more practical meaning: an ontology is the agreed set of definitions that lets different people, teams, and software talk about the same things in the same way, without ambiguity.

What does an ontology contain?

Most ontologies are built from a small number of recurring ingredients:

• →

  Classes (or types): the kinds of thing that can exist, such as a Person, Company, or Invoice. In HASH these are expressed as entity types.

• →

  Properties (or attributes): the pieces of information a thing is expected to carry, such as a person’s Name or Date of Birth. See properties and data types.

• →

  Relationships (or links): the ways things can be connected to one another, such as a Person being employed by a Company. See links.

• →

  Hierarchy and inheritance: more specific types can inherit the expected properties of more general parent types (for example, a Customer is a kind of Person).

• →

  Axioms and constraints: rules that further restrict how the concepts may be combined, enabling automated reasoning and validation.

How are ontologies used?

Ontologies are the backbone of the semantic web and of most knowledge graphs. By committing to a shared ontology, otherwise siloed datasets can be linked together, queried consistently, and reasoned over. Standards such as RDFS and OWL (the Web Ontology Language) provide widely-used formats for publishing ontologies, while domain-specific ontologies exist across fields as varied as healthcare, finance, and manufacturing.

A well-designed ontology delivers several benefits:

• →

  Interoperability: data from different sources can be combined because everyone refers to the same definitions.

• →

  Consistency: information is captured in a structured, validated way, reducing duplication and error.

• →

  Inference: because relationships and constraints are explicit, software can derive new facts that were not stated directly.

Ontologies in HASH

In HASH, the collection of types you create — entity types, link types, property types, and data types — together form a living ontology of the domain you care about. Unlike traditional ontologies, which often require every participant to agree on a single global vocabulary up front, HASH is multi-tenant and versioned: each user or organization can define and extend types in the terms that matter to them, while remaining interoperable with the wider graph of knowledge. This makes it possible to evolve an ontology over time as understanding of a domain matures, without breaking the data already described by earlier versions.