# Ontology

An ontology formally defines the concepts in a domain, their properties, and the relationships and rules that connect them.

- Type: framework
- Domain: Experience & interface
- Level: Advanced
- Last reviewed: 2026-09-21
- Original: [Soroush’s Product Wiki](<https://wiki.soroush.uk/topic/ontology/>)

## What it means

An ontology is an explicit model of a domain that describes classes of things, their attributes, their relationships, and sometimes constraints that allow new facts to be inferred. Unlike a simple taxonomy, which mainly classifies concepts, an ontology can represent richer statements such as a clinician treating a patient at a facility or a product being compatible with another product.

## Why it matters

Ontologies make meaning more consistent across systems and teams. They can support semantic search, knowledge graphs, data integration, recommendation, and machine reasoning by clarifying when differently named data refers to the same concept. The modeling work also exposes hidden assumptions that otherwise remain embedded in databases and interfaces.

## When to use it

Use an ontology when the product must combine complex data from several sources, represent many relationship types, or support queries that a category tree cannot answer. It is appropriate for domains such as healthcare, science, commerce, and asset management where shared semantics justify ongoing specialist governance.

## How it works

1. Specify the domain, competency questions, intended users, and level of inference the model must support.
2. Reuse established vocabularies where practical before creating new classes and properties.
3. Model core entities, attributes, relationships, identifiers, and constraints with domain experts and data practitioners.
4. Create representative instances and test whether the ontology can answer the competency questions without awkward exceptions.
5. Publish documentation, ownership, mappings, and versioning rules so dependent systems can adapt safely.

## A practical example

A cultural heritage platform models creators, works, physical objects, exhibitions, places, and time periods as distinct entities. A painting can be created by a person, held by an institution, shown in several exhibitions, and associated with a movement. Users can then explore relationships that a single folder hierarchy would hide.

## Where it breaks

Ontology work can become expensive and abstract, especially when teams model every theoretical distinction before proving a user need. Domain experts may disagree about definitions, and inference can produce surprising results when rules are incomplete. A lightweight schema or taxonomy is often better when the required relationships are simple.

## Common mistakes

- Treating an ontology as a more elaborate navigation tree.
- Modeling the entire domain before validating a small set of valuable questions.
- Creating local concepts when stable, interoperable vocabularies already exist.

## Visualization

[View the visualization on the website](<https://wiki.soroush.uk/topic/ontology/#visualization-ontology>)

## Sources

1. [OWL Web Ontology Language Overview](<https://www.w3.org/TR/owl-features/>) — World Wide Web Consortium · W3C specification overview
2. [OWL: Semantic Web Standards](<https://www.w3.org/OWL/>) — World Wide Web Consortium · Standards hub
3. [OWL 2 Web Ontology Language Document Overview](<https://www.w3.org/TR/owl2-overview/>) — World Wide Web Consortium · W3C recommendation

## Related topics

- [Taxonomy](<https://wiki.soroush.uk/topic/taxonomy/>)
- [Object-Oriented UX](<https://wiki.soroush.uk/topic/object-oriented-ux/>)
- [Search Design](<https://wiki.soroush.uk/topic/search-design/>)
