Drawing hierarchical structures in a way that clearly represents the information they carry is a centuries-old problem that artists have dealt with. Today, the widespread use of ontologies requires this problem to be viewed from a new perspective: not only is it a drawing problem, but it requires tools for exploring, changing, and querying. Different tools are available today, and choosing which to use is not a trivial task.
As indicated by the title, this paper is a broad survey of the ontology visualization problem. A major part of the paper presents the available tools for the visualization of ontologies. Twenty-four are illustrated, and some are compared in a table. They use as visualization methods indented lists, node-link trees, and zoom and focus; moreover, they offer tools for building and querying ontologies. There is no mention of cognitive studies to put these features into perspective.
The authors also wanted to determine the best features for users with basic domain knowledge, using ontologies as “browsing aids.” Three tools, related to Protégé, were selected: DLQuery (a tool that facilitates the definition and checking of classes), OntoGraf (a tool offering various layouts), and OWL2Query (a system that facilitates queries in SPARQL). The authors briefly describe the design of an experiment for evaluating user satisfaction with the cognitive support provided by the three systems in using an ontology containing 15 classes and about 2,600 instances. A group of people with minimal training received a set of five queries to answer from the ontology. The results are shown in tables showing TaskLoadIndex (TLX) and time needed. People were able to more easily and effectively use OntoGraf. Why? Which features make it so user friendly? The analysis stops here.
As the authors conclude, these results do not say which system is the best. The value and the acceptance of a tool, we may conclude, depend on the size of the ontology, the previous knowledge of the user, the stability of the domain knowledge, and the task to accomplish; these are too many parameters for a simple experiment.