Can Ontologies Systematically Help in the Design of Domain-Specific Visual Languages?

Maria das Graças da Silva Teixeira, Ricardo de Almeida Falbo, and Giancarlo Guizzardi

Abstract. There has been a growing interest in Domain-Specific Visual Modeling Languages (DSVL) and their support for domain understanding and communication. However, the quality of these languages fundamentally depends on how well their structure reflects the structure of the abstractions constituting the underlying domain conceptualization. Since a well-founded domain ontology aims at faithfully representing a domain, it can be seen the ideal input for engineering a DSVL. In this paper, we present an experiment that analyses the performance of computer students in interpreting instance models by varying the concrete syntax of the language used. We contrast a generic notation (UML-based notation for object diagrams) and a domain specific notation that was designed based on a well-founded ontology for the domain of organizational structures. The hypothesis is that the performance of participants in interpreting the models using the domain specific notation is better than those who do it through a generic notation. Performance is evaluated by taking response time and correctness of the answers into account. The results confirm, but also contradict the hypothesis initially formulated.

Keywords: Domain-Specific Visual Language, Ontology, Empirical Study, Conceptual Modeling

LNCS 8185, p. 737 ff.

Full article in PDF | BibTeX

lncs@springer.com
© Springer-Verlag Berlin Heidelberg 2013