Ontologies are an increasingly important method of knowledge representation, especially in the arena of the Semantic Web. In this book, David Stuart aims to demonstrate their importance, show how information professionals can make an important contribution to their development, and provide a practical introduction to the development process. He begins by defining an ontology, and moves on to cover the Semantic Web, existing ontologies and their adoption, how to build an ontology, and interrogating ontologies. The final chapter looks at the future of ontologies and the role of the information professional.
There is a great deal of helpful introductory material here for students and professionals who have little, if any, background in this area. The examples, in a number of cases, are taken from the library world and include such things as FRBRoo and the Bibliographic Ontology. Stuart is persuasive in arguing that information professionals should get involved in developing and maintaining ontologies, as well as knowing how to make use of them. He makes some important points about the distinction between using ontologies as knowledge discovery tools and using ontologies for knowledge representation – though he could have taken more notice of the semantic issues around the philosophical validity of ontologies.
Given the inconsistent use of terminology in this field, defining an ‘ontology’ is probably an inevitable starting point, if a rather dry one. Stuart offers a broad definition: ‘an ontology is a formal representation of knowledge with rich semantic relationships’ (p. 12). But he adds his own inconsistencies. At times, he talks about vocabularies when he seems to mean ontologies, and vice versa. Is schema.org a vocabulary or an ontology? In what sense is Dublin Core an ontology? He also tends to elide the difference between an ontology and an ontology language. RDF, SKOS and OWL are described as ontologies, but they are, strictly speaking, ways of expressing ontologies through formal notation rather than ontologies in the true sense.
The book also includes a lot of material which is tangential to ontologies: Linked Data, controlled vocabularies, thesauri, natural language processing, named entity recognition, knowledge graphs, and so on. This is very valuable for putting the use of ontologies into a wider context, but beginners might find the resulting complexity quite challenging, especially since some of the diagrams could have been much clearer. The relationship between all these elements might also have been made clearer by discussing some detailed examples of existing digital services in different disciplines which use ontologies in combination with these other tools, such as Pelagios for ancient history. Overall – this is a good place to start when grappling with ontologies in our profession.
University of Western Australia
[email protected]
© 2018 Toby Burrows
https://doi.org/10.1080/24750158.2018.1430453