Skip to content

Annotations

Jump to bottom
chris-day edited this page Nov 21, 2025 · 9 revisions

Annotations – Complete Specification

Annotations in uml2semantics provide a structured mechanism for attaching governance metadata, documentation, quality indicators, provenance, lifecycle information, stewardship roles, and ISO-aligned semantic labels to ontology entities.

The annotation system is fully declarative and driven by two TSV files:

  1. AnnotationProperties.tsv – defines annotation vocabulary
  2. Annotations.tsv – applies annotation assertions to ontology elements

Both files integrate seamlessly with classes, properties, datatypes, individuals, and the ontology itself.

1. Purpose of the Annotation System

Annotations allow you to enrich your ontology with metadata such as:

  • natural language descriptions
  • governance tags (e.g. steward, owner, version, status)
  • provenance (e.g. source, importedFrom, creation date)
  • quality information
  • FAIR metadata
  • controlled vocabulary links
  • lifecycle states (draft, approved, deprecated)

uml2semantics ensures annotations are correctly serialised as OWL 2 annotation assertions in Turtle, RDF/XML, or JSON-LD.

2. AnnotationProperties.tsv – Defining Annotation Vocabulary

This TSV defines the annotation properties that can be used across the model.

2.1 Required Columns

Column Description
Curie CURIE for the annotation property, e.g. dct:creator
Name Human-readable label
Definition Description of the annotation property and its intended usage

2.2 Example

Curie           Name         Definition
dct:creator     Creator      The person or system that created the entity.
dct:source      Source       Source data or upstream artefact.
iso:status      Status       Governance lifecycle label.

uml2semantics generates OWL axioms of the form:

AnnotationProperty: dct:creator
  Annotations: rdfs:label "Creator"

3. Annotations.tsv – Annotation Assertions

This TSV applies annotation assertions to classes, properties, datatypes, individuals, or the ontology itself.

3.1 Required Columns

Column Description
TargetCurie CURIE of entity being annotated
AnnotationProperty CURIE referencing AnnotationProperties.tsv
Value Literal string or IRI
Language Optional language tag (ISO 639‑1)
Datatype Optional literal datatype

3.2 Value Interpretation Rules

  • If Datatype is given, Value is treated as a typed literal.
  • If Language is given, Value is serialized as a language-tagged literal.
  • If neither is provided, Value is a plain string literal.
  • If Value begins with a registered prefix, it is interpreted as an IRI.

Examples:

Value Result
"Account Transactions" plain literal
"Account" with language en "Account"@en
"2024-01-20" with datatype xsd:date "2024-01-20"^^xsd:date
skos:Concept123 IRI reference

4. Example Annotation Usage

Example Rows

TargetCurie    AnnotationProperty   Value                       Language   Datatype
iso:Account    rdfs:label           "Account"                   en
iso:Account    dct:source           "ISO 20022 Logical Model"
iso:Amount     iso:status           "approved"
iso:Amount     dct:creator          "Chris Day"
iso:Amount     dct:modified         "2025-02-10"                ""        xsd:date
iso:GBP        skos:prefLabel       "British Pound Sterling"    en

Resulting OWL (Turtle)

iso:Account rdfs:label "Account"@en .
iso:Account dct:source "ISO 20022 Logical Model" .
iso:Amount iso:status "approved" .
iso:Amount dct:creator "Chris Day" .
iso:Amount dct:modified "2025-02-10"^^xsd:date .
iso:GBP skos:prefLabel "British Pound Sterling"@en .

5. Governance Metadata Examples

Annotations are ideal for recording:

5.1 Status Indicators

iso:status = draft | proposed | approved | deprecated

5.2 Stewardship

gov:steward = “Data Architecture Team”
gov:owner   = “Payments Domain Owner”

5.3 Provenance and Versioning

dct:source = “ISO 20022 2013 Edition”
dct:modified = “2024-10-02”
owl:versionInfo = “v2.3.1”

5.4 FAIR Metadata

fair:category   = “Interoperability”
fair:score      = “0.83”

6. Annotation Use Cases

Use Case 1 – Semantic Documentation

Annotate entities with descriptive labels, domain definitions, glossary references.

Use Case 2 – Data Governance

Track ownership, approval workflow, status, lifecycle.

Use Case 3 – Provenance & Lineage

Record upstream systems, data sources, transformation authors.

Use Case 4 – Compliance Metadata

Mark controlled classes, explain regulatory requirements.

Use Case 5 – FAIR Enablement

Provide discovery, accessibility, interoperability and reusability metadata.

7. Best Practices

  • Maintain a curated list of annotation properties.
  • Use one TSV row per annotation.
  • Prefer CURIE values over hardcoded IRIs.
  • Use typed literals for dates, numbers, and formal values.
  • Include natural language strings for human readability.

8. Mermaid Diagram – Annotation Workflow

graph TD
  AP[AnnotationProperties.tsv] --> P[Annotation Properties]
  AN[Annotations.tsv] --> A[Annotation Assertions]
  A --> O[Ontology Elements]
Loading

9. Next Steps

Continue to:

Return to Home.

Clone this wiki locally