Ontology in software engineering

Abstract

During the past years, ontological thinking and design have become more and more popular in the field of Artificial Intelligence (AI). More recently, Software Engineering (SE) has evolved towards more conceptual approaches based on the extensive adoption of models and meta-models. This paper briefly discusses the role of ontologies in SE according to a perspective that closely matches the theoretical life-cycle. These roles vary considerably across the development lifecycle. The use of ontologies to improve SE development activities is still relatively new (2000 onward), but it is definitely no more a novelty. Indeed, the role of such structures is well consolidated in certain SE aspects, such as requirement engineering. On the other hand, despite their well-known potential as knowledge representation mechanisms, ontologies are not completely exploited in the area of SE. We first (i) proposes a brief overview of ontologies and their current understanding within the Semantic Web with a focus on the benefits provided; then, the role that ontologies play in the more specific context of SE is addressed (ii); finally, we deal with (iii) some brief considerations looking at specific types of software architecture, such as Multi-Agent Systems (MAS) and Service-Oriented Architecture (SOA). The main limitation of our research is that we are focusing on traditional developments, where phases occur mostly sequentially. However, industry has fully embraced agile developments. It is unclear that agile practitioners are willing to adopt ontologies as a tool, unless we ensure that they can provide a clear benefit and they be used in a lean way, without introducing significant overhead to the agile development process.