Ontology
Give records shared business meaning and connect that meaning to queryable values
An ontology defines the business things a project talks about, such as Equipment, and the properties and relationships describing them. It separates what a term means from where its current values come from.
The pieces
| Piece | Purpose | Equipment example |
|---|---|---|
| Definition | Versioned classes, properties, relationships and shapes | Equipment has an equipment identifier and temperature |
| Binding | Connects a term to a source and read method | Temperature comes from a selected endpoint column |
| Materialized fact | A mapped entity/property/relationship persisted through a fact store | P-101 has a stored temperature observation |
| Assertion | Reviewable statement with its own evidence and lifecycle | An incident supports a claim that P-101 failed |
| Published query | Released input/output contract over the model and bindings | Return measurements for one equipment identifier |
Definitions, compiled packages, bindings and queries belong to the project. Plugin installations and endpoints are workspace resources the project references. A definition asset/version is different from the compiled package/version used by bindings and mappings.
Reading values
Virtual bindings read a supported endpoint at query time. Materialized bindings read persisted ontology facts. Asserted bindings select assertion values according to their resolution policy. Inferred bindings compute supported expressions from other terms. One query can combine methods when entity identity and its contract align.
Creating Equipment in a model does not create equipment records. A model graph shows structure, not proof that data was read or a materialization run succeeded. Publication stabilizes query logic; live source values can still change.
Example: P-101
A workspace endpoint selects a measurements table. The project defines Equipment and binds equipmentId/temperature to source fields. A query matches equipmentId against its subject input and returns P-101's measurement. The entity identifier, business identifier, source column and query output name are distinct unless explicitly mapped.
Learn the full lifecycle
Ontology covers definitions, bindings, materialization, assertions and queries. Your first model and query contains a complete fixture. Use assertions and evidence for recorded claims rather than treating every statement as an ordinary source fact.