Advanced ontology visualization with OWLGrEd Jānis Bārzdiņš, Kārlis Čerāns, Renārs Liepiņš, Artūrs Sproģis Institute of Mathematics and Computer Science, University of Latvia, Raina blvd. 29, LV-1459, Riga, Latvia Janis.Barzdins@lumii.lv, Karlis.Cerans@lumii.lv, Renars.Liepins@lumii.lv, Arturs.Sprogis@lumii.lv Abstract. Intuitive ontology visualization is a key for their learning, exchange, as well as their usage in conceptual modeling. OWLGrEd is a visual tool for compact graphical UML-style rendering and editing of OWL 2.0 ontologies. Here we introduce into OWLGrEd visualization annotations that allow defining custom visual presentations of ontology entities on the basis of annotations attached to these entities. The introduced mechanism of attaching visual annotations to annotation properties used in the ontology, besides being convenient for attaching graphical shape to user-defined annotations, appears to be powerful enough to cover visualization of e.g. UML stereotypes and other UML constructs that do not have direct counterpart in the "logic" part of OWL. Keywords: OWL ontologies, visualization, annotations, OWLGrEd. 1 Introduction Intuitive ontology visualization is a key for their learning, exchange, as well as their usage in conceptual modeling. There are a number of tools (e.g. [1,2,3,4]) for rende- ring and/or editing OWL ontologies in a graphical form. ODM [1] and OWLGrEd [4] are oriented on visualization of OWL ontologies in the style of UML [5] class diagrams. The main idea of OWLGrEd has been to provide compact graphical notation for OWL ontologies, by combining the graphical visualization facilities of UML class diagrams with textual facilities of OWL Manchester encoding [6]. Although OWLGrEd can be successfully used for ontology presentation and editing in a UML-style graphical form (cf. [7]), the practical task of ontology-based modeling suggests further facilities that would be welcome in a graphical ontology editor. The ontology designer may attach certain specific meaning to some ontology entities by user-defined annotation properties and then may wish to create specific visualization patterns for the introduced properties. For instance, if an annotation assertion relates an OWL class to a database table expression, a database icon besides the table expression field may be appropriate in rendering this annotation assertion. There are also natural modeling constructs in UML that do not have direct counterpart in OWL, such as stereotypes, composition relation and derived union of properties (or “abstract” property A, not having other subject-object pairs than those of union of B and C). The visual annotation framework that we introduce in this paper allows both visual effect specification for user-defined annotations and visual modeling of the mentioned “advanced” UML concepts (these concepts may be said to correspond to “user-defined” annotations from OWL perspective, as well). In what follows, after brief reviewing of basic OWLGrEd ontology modeling facilities we proceed to the main point of outlining the visual annotation framework and some its usage examples, including advanced UML construct modeling. 2 Basic OWLGrEd constructs OWLGrEd provides graphical notation for OWL, based on UML class diagrams. We show OWL classes as UML classes, data properties as class attributes, object properties as associations, individuals as objects, etc. We enrich the UML class diagrams with the following new extension notations (cf. [4, 8]): • fields in classes for equivalent class, superclass and disjoint class expressions written in Manchester OWL syntax [6]; • fields in associations and attributes for equivalent, disjoint and super properties and fields for property characteristics, e.g., functional, transitive, etc.; • anonymous classes containing equivalent class expression but no name; • connectors for visualizing disjoint, equivalent, etc. axioms; • boxes with connectors for n-ary disjoint, equivalent, etc. axioms; • connectors for visualizing object property restrictions some, only, exactly, etc. OWLGrEd provides option to specify class expressions in compact textual form rather than using separate graphical element for each logical item within class expression. If an expression is referenced in multiple places, it can optionally be shown as an anonymous class. An anonymous class is also used as a base for property domain/range specification, if this domain/range is not a class. Thing {owl} eats eaten-by-animal is-part-of eaten-by {> Herbivore Carnivore = Animal eaten-by some [1..*] Tree Tasty-plant = Animal and (eats only and (eats some Animal) (Plant is-part-of or (is-part-of only Plant))) only <