Reviews

Newsletter Issue: 
Vol 37 no. 4 December 2017

Richard N. Leigh, Column Editor
 

Practical Ontologies for Information Professionals
by David Stuart

 

Practical Ontologies for Information Professionals is a lucid and forceful argument for the importance of ontologies in navigating information overload, and the critical role information professionals can play in both creating new ontologies and developing existing ones. The author defines ontologies as “formal representations of knowledge with rich semantic relationships” (1), especially those intended for recoding in linked data for placement on the semantic web. Librarians may be more likely to refer to these knowledge organization systems as “controlled vocabularies”, “taxonomies”, or “thesauri”, depending on their specialty. Catalogers are already familiar with several examples of ontologies, of course, including subject headings and authority files.

Chapter 1 (“What is an ontology?”) focuses on how ontologies can be used by libraries and digital repositories. The author shows how ontologies provide indexing support, retrieval support, and organization and navigation support. Information encoded in an ontology can be queried in complex ways : to highlight relationships, to show hierarchies, to visualize data, etc.

Chapter 2 (“Ontologies and the semantic web”) shows how ontologies are currently being used in libraries, cultural heritage institutions, and commercial organizations. The author explains the potential benefits of “moving from a web of documents to a web of data, from one that is primarily designed to be read by humans to one that can be read by machines” (27). The Resource Description Framework, RDF triples, and XML are discussed, as are classes, subclasses, and properties.

Chapter 3 (“Existing ontologies”) lists the most widely-used ontologies, comparing their respective strengths and weaknesses. Ontologies are considered for constructing ontologies (RDF, RDFS, SKOS, OWL2), for usage in libraries (Dublin Core, Bibliographic Ontology, FRBR, RDA, Bible Ontology), for usage in cultural heritage institutions (Europeana Data Model, CIDCO-Conceptual Reference Model), and for usage on the web (DBpedia ontology, Friend of a Friend ontology, Schema.org, Facebook Open Graph Protocol).

Chapter 4 (“Adopting ontologies”) explains how to select an ontology for usage, based on variables like clarity, coherence, extendibility, minimal encoding bias, minimal ontological commitment, goodness of fit, format, currency, adoption, documentation, licensing, etc. Benefits of choosing an extant ontology include cost-effectiveness, colleagues to communicate with, and facilitating the interoperability of data. The author suggests surveying the available options in an ontology library (such as Linked Open Vocabularies, BARTCO.org, The Taxonomy Warehouse, BioPortal) or an ontology search engine (like Swoogle, Watson, Falcons).

Chapter 5 (“Building ontologies”) describes how to develop an ontology from scratch, via a 12-step methodology devised by the author (with inspiration from four prominent ontology development methodologies originally published between 1995-2001). Stuart also provides a case study : developing a Bibliometric Metrics Ontology element set with Protégé - a free, open-source ontology editor and framework created by Stanford University. 74 | Page

Chapter 6 (“Interrogating ontologies”) focuses on how to search semantic web ontologies, via Simple Protocol and RDF Query Language (SPARQL) or web crawlers. Ontologies may be queried not only for informational purposes (i.e. research), but also to determine if the ontology is suitable for repurposing or to understand how the ontology is currently used. Simply reading the documentation is not advisable for a large or complex ontology; software or online tools will be better, especially if utilized by an experienced computer scientist.

Chapter 7 (“The future of ontologies and the information professional”) predicts possible developments in the usage of ontologies and suggests possible avenues for librarians to create their own ontologies. Natural Language Processing and machine learning are discussed, both for their own merits and as part of a hybrid approach involving human vetting of their results. The author predicts ever-increasing diversity for ontologies - open and closed, formal and informal, centralized and distributed, etc. Practical Ontologies for Information Professionals is extremely well-organized, allowing it to be read either straight-thru or referred to for ready-reference. Each chapter has an easily understandable title and equally informative subsection titles (ex : “Alternative semantic visions”). Even the figures and tables are well-described! The prose is admirably clear, too, especially for such a complicated topic. Anything that is simultaneously both this abstract and this technical will inevitably be a challenging read, but the author does an excellent job of helping readers scale the “knowledge pyramid” from data and information to knowledge and wisdom.

Published in 2016 by: Neal-Schuman, Chicago; Facet Publishing, London (viii, 184 p. ; 24 cm.) ISBN 978-0- 8389-1511-0 (U.S.); 978-1-78330-062-4 (U.K.); (softcover : $79.20, £59.95)

Reviewed by: Richard N. Leigh
Metadata & Digital Resources Developer
University Libraries
Ball State University