We present an extension to the OWLGrEd visual ontology editor that enables visual annotation of OWL data and object properties, and instance property assertions with contextual qualifiers such as temporal scope, provenance, and confidence. Inspired by RDF 1.2 reified triples, our approach uses OWL 2 annotation mechanisms to enrich both property declarations and assertions without altering ontology semantics. This results in an intuitive visual representation of contextual data in OWL ontologies.
O
A wide range of ontology visualization tools provide graphical representations of OWL 2 [
In this work, we present an extension to the OWLGrEd ontology editor that fills this gap by enabling visual qualifiers on relationships. Our approach enables ontology developers to attach contextual data (e.g., time intervals, source identifiers, trust/confidence values) to both property declarations (schema-level relationships) and individual property assertions (instance-level facts) within an OWL ontology. This is achieved using the OWL 2 annotation mechanisms, which enable the linking of additional information to ontology entities and axioms. By leveraging standard annotation properties in OWL 2, we ensure that qualifiers are encoded in a semantically compatible way – they do not alter the formal semantics of the ontology, but enrich it with supporting information. The result allows users to explore not just the structure of an ontology but also the context of its relationships within a single integrated diagram.
OWLGrEd [
CEUR
ceur-ws.org
axioms such as disjointness or equivalence. OWLGrEd has been built in GRTP visual tool development
platform [
To introduce property qualifiers into the OWLGrEd visual notation, we provide the option to visually specify object and data properties using dedicated graphical boxes, and enable describing qualifiers within compartments of these property boxes.
At the instance level, we provide a visual notation for recording any annotations on data and object property assertions, with the intention to use them with the annotation properties introduced as data and object property qualifiers on the schema level.
We use color-coding to distinguish property types: object properties and object property assertions are depicted in blue hexagonal boxes, while data properties and data property assertions use green hexagonal boxes (see Figure 2). Each box can graphically connect to relevant ontology elements: classes in the case of property declarations, or individuals in the case of assertions. The graphical connections follow the logical structure of RDF triples: the subject (either a class or an individual) links to the property box with an outgoing arrow, and the object (class or individual) connects via an incoming arrow. This left-to-right flow visually mirrors RDF’s subject–predicate–object structure and supports intuitive reading of both ontology declarations and instance-level assertions. Alternatively, information about domain and range (in the case of declarations) or subject and object/value (in the case of assertions) can be displayed textually within the box itself using labelled fields such as “ Domain”, “Range” (see the third row of Figure 2), and “Value” (fourth row of Figure 2). This dual representation supports both compact and detailed modelling preferences.
Additionally, generalization relationships can be used with object and data properties to visually encode sub-property axioms, clearly representing hierarchical relationships among properties within the ontology (as shown in the first row of Figure 2).
Each of the property and assertion visual elements can be enriched with qualifiers – fields that provide contextual information about the axiom, such as temporal validity, certainty, or provenance. Qualifiers like datePaid, mark, and tutor are displayed as labelled fields within the hexagonal box.
OWL 2 does not have a native notation specifically dedicated for expressing property qualifiers. However, it provides mechanisms, such as annotation properties, that can efectively represent this information. In this work, we use the OWL 2 annotation mechanism to introduce property qualifiers such as temporal scopes, provenance information, and confidence levels. These annotations enrich both property declarations (schema-level relationships) and property assertions (instance-level facts) without modifying the formal semantics of the ontology.
The property qualifiers (e.g., dateTo, tutor ) are encoded in the ontology as annotation properties. These qualifiers are attached to the corresponding data or object properties using a dedicated annotation property ex:qualifier 1. The expected type of qualifier values is specified using a nested annotation with annotation property rdfs:range, which may refer to a datatype or a class IRI, depending on whether the qualifier values are literals or resources (see the example of the property :salary below). Additional information, such as cardinality constraints (e.g., owl:maxCardinality annotation property) indicating how many times a qualifier may or must appear, can also be included. For example, the following axiom indicates that the data property :salary is qualified by :dateTo, which is expected to be a datetime and not occur more than once per salary assertion: AnnotationAssertion(
Annotation(rdfs:range xsd:dateTime) Annotation(owl:maxCardinality "1"^^xsd:nonNegativeInteger) ex:qualifier :salary :dateTo)
Once defined and annotated in this way, the same annotation properties are consistently applied to annotate the corresponding property assertions on the instance level, ensuring uniform usage across the ontology. For example, a data property assertion like :Dave :salary “3000” can be annotated with qualifier dateFrom and value “2023-01-06 ”.
We have chosen to use the local semantics for the qualifier visualizations as the qualifier range (and cardinality) assertions are added directly into the axioms asserting a particular qualifier for a property (e.g., AnnotationAssertion(ex:qualifier :salary :dateTo) in Figure 2). Another option would have been to introduce the qualifier property characteristics globally, as follows: AnnotationAssertion(ex:qualifier :salary :dateTo) AnnotationPropertyRange(:dateTo xsd:dateTime) AnnotationAssertion(owl:maxCardinality :dateTo 1)
The local option has been chosen since it can be expected that the same qualifiers could be applied in diferent places in the ontology, together with a diferent characteristics.
Although these qualifier annotations do not afect the ontology’s formal semantics from an OWL reasoning perspective, they enrich axioms with contextual information. This enhances visualization and better readability for users while maintaining full compatibility with OWL 2 standards.
This approach is conceptually inspired by the RDF 1.2 notion of reified triples [
We have implemented a prototype extension of the OWLGrEd editor to support visual annotation of OWL axioms.2 The prototype includes a sample mini-University project using the extra annotation notation described in this paper.
We have extended the UML-style OWL diagramming tool OWLGrEd with the ability to represent qualifiers on both schema and instance relationships, using a clear syntax that extends the graphical notation without overcomplicating it. This extension introduces new diagram elements and syntax for qualifiers while remaining fully compliant with OWL 2 standards (qualifiers are stored as OWL annotations, and thus preserve semantic compatibility).
We expect that our contributions would enhance the usability of the tool - ontology engineers and domain experts can now capture and view contextual data directly in the ontology diagram, without needing external documentation. The integrated qualifiers not only improve the expressiveness and understandability of ontology visualizations, but also open the door for some forms of enhanced reasoning and analysis: since the context qualifiers are formally represented, they can be used by reasoning tools or queries (e.g., temporal reasoner) to provide richer answers and explanations.
As future work, we plan to implement the described graphical notation in a web-based environment that supports the full lifecycle of visual ontology development. We also intend to conduct a detailed evaluation of the notation within the context of the upcoming ontology development tool.
This work was supported by activity 1.1.1.9 Research application No 1.1.1.9/LZP/1/24/037 of the Activity 2https://owlgred.lumii.lv/downloads/OWLGrEd_extended.zip “Post-doctoral Research” “Visual methods and tools for ontology management”, and the Latvian Council of Science, project “What’s in Your Knowledge Graph? ” (Project No. lzp-2024/1-0665).
During the preparation of this work, the authors used ChatGPT (GPT-4) in order to: Improve writing style; Grammar and spelling check. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the publication’s content.