Ontologically-based context checking in arbitrary programming languages

Abstract

Ontologies provide an explicit representation of concepts and their relationships, which provides a strong context for the use and manipulation of data within a problem domain. If the data is referred to without the ontological framework, any association with its original context is lost. A complex system may be implemented with several, different, programming languages and, while an ontological description may be in use at its core, data transferred into or out of this core may traverse many other programming language semantics that do not have the expressiveness or logical foundations of an ontological language. As a result, programming errors may arise, due to the accidental use of one value where another was expected, or through an erroneous or incomplete understanding of the type of a given value. Rather than expect a system to be implemented completely within an ontological framework, with associated computational overhead, we propose an approach where the ontology is used to produce a pre-processing mechanism that, when used in conjunction with traditional compilation tools, can reduce the misuse of ontologically-defined terms