Towards a Theory on the Role of Ontologies in Software Engineering Problem Solving: Conclusions from a Theoretical Model of Methodological Works

Abstract

We present and validate a theoretical model of methodological works in Software Engineering that, without claiming for completeness, allows us to investigate the role of ontologies in the problem solving process related with the development of software. Our main conclusion is the potential of ontologies as resources for an individual to think during problem solving. We argument that suitable ontologies can support solving strategies as well as motivate their invention. We also conclude the importance of accompany an ontology with knowledge that guides the engineer in reasoning with its concepts. The model regards a methodological work as an heterogeneous theory about a class of problems and about a number of conceptual elements. Some of the elements are ontologies, which play the role of identifying and relating aspects of the knowledge about the class of problems, making up novel perspectives on the problems that may promote solving strategies. For illustration purposes, we take Jackson’s “Problem Frames” as a case study. We analyse this work through the former model, identifying the ontologies, guides, and promoted strategies. Then we propose an alternative ontology, based on that used in the KAOS approach; we reformulate some parts of Jackson’s work through this ontology and propose a strategy as well as some guides.