Investigations in ontology development : the use of standard office tools and Arabic script

Abstract

Ontologies are useful models for representing knowledge, especially in complex and dynamic domains such as biology or medicine. However, capturing such complex knowledge is essential to ensure the accuracy and reliability of any ontology. Therefore, building an ontology is a collaborative process which involves different communities who are required to work effectively in developing it; this includes domain specialists who understand the knowledge and ontology developers who identify the ontological representation of that specific knowledge. This collaboration requires a variety of instruments as each community has its own tooling and techniques. In order to facilitate the interaction between domain specialists and ontologists, a documentcentric workflow focused on the use of English and standard office tooling is designed. The aspiration here is to introduce a framework for building ontologies that allows both parties to use their own toolchains and for those toolchains to work together. Within biology and medicine, probably the most common tools for manipulating and sharing data are tools such as Microsoft Excel spreadsheets and Word documents. Unfortunately, these environments are far distant from the formal structured representation of the ontology development environments with which the ontologists work. This research investigates approaches to applying standard office tools for developing and presenting ontologies, demonstrates how a narrative structure can be applied to an ontology, and finally, discusses the opportunities and challenges associated with internationalising an ontology development environment into Arabic. The goal is to bridge the gap, both in tooling and language, between ontological experts and biologists. The results showed that it is possible to rethink ontology tooling and how various ontology representations are used in the development process. It has been demonstrated that office tooling, used naturally, can be integrated into the source code of an ontology, that it is possible to add narrative documentation to an ontology, and finally, to represent that narrative structure in Word, an application that most domain users are intimately familiar with. Furthermore, the user evaluation has shown that users can interact well with this form of representation. Taken together, these mechanisms suggest a plethora of future directions for developing ontologies and, more broadly, computational representations of complex knowledge domains.