Extracting ontologies from legacy systems for understanding and re-engineering

Abstract

Ontology has been investigated in the context of knowledge sharing among heterogeneous and disparate database and knowledge base systems. Our recent study and experiments suggest that ontology also have a great potential for legacy software understanding and re-engineering. In this paper we consider an ontology to be composed of four elements: classes, relations, functions and instances. We show these four elements forming an ontology for a legacy system can be extracted from the code of the concerned system using the existing software re-engineering tools. We then present our vision how the obtained ontology can be applied to understanding and eventually better re-engineering the legacy systems.