Authors: Grigoris Antoniou & Frank van Harmeten
Publisher: MIT Press, 2004
ISBN 0262012103
This quarter’s issue of The Data Administration Newsletter includes part two of a three-part series I am publishing to describe how the world of data modeling links to the newly discovered
(by us data modelers) world of ontological engineering. After an introduction to the topics last quarter, this quarter begins the process of converting a sample data model to the ontology language
OWL.
This has been a very difficult series to write, not the least because, while there is a substantial body of knowledge describing both OWL and ontological engineering in general, it comes out of the
academic world and is a little intense for the casual data modeler to make sense out of.
Semantics and Ontology were originally two schools of greek philosophy. Ontology was the study of what exists, and semantics was the study of how to describe it. These have been given new life in
the last twenty years or so in the artificial intelligence community, but they have only really “gone public” in recent years. This has been triggered by the general recognition that
different departments in an organization (and their systems) don’t relaly communicate very well and could do with a better understanding of their semantics, plus the specific advent of
“The Semantic Web”.
The Semantic Web is the brainchild of Tim Berners-Lee, the inventor of the World-wide Web. It is his vision that not only should documents be available readily, but the semantics of their content
should be accessible. This has prompted the development of languages like OWL, a set of XML tags to describe the semantics of a document.
Now, however, there is a reasonable introduction to the subject! A Semantic Web Primer, by Grigorious Antoniou and Frank van Harmeten, is an excellent description for an intelligent reader who is
not deeply into artificial intelligence, linguistics and semantics. This is the book I wish I’d had when I started the project that led to the TDAN articles.
After reading several books on the Semantic Web, I can say with confidence that this is by far the clearest and easiest to understand. Since the ontological languages of RDF, RDF Schema, and OWL
are typically represented in XML, the book begins with a section describing XML to the uninitiated, which includes an excellent description of XML Schema—the alternative to document type
definitions (DTDs). (As it happens, the RDF and OWL languages are simply examples of XML Schemas.)
The Resource Description Framework (RDF) is a way of parsing sentences that has been around for a long time. This lets you identify subject, predicate, and object. What it does not do is address
the distinction between instances and classes. RDF Schema corrects that shortcoming by adding the concepts of class and instance. OWL then allows you to describe constraints. These definitions
sound very reasonable, yes? Most books on this subject are not nearly as easy to understand as this. A Semantic Web Primer is.
The tutorials on RDF, RDF Schema and OWL are the clearest I have yet encountered. The descriptions not only cover the logical aspects of the languages, but they also are very good at dealing with
some of their quirks. Recognizing that they have quirks goes a long way toward clarifying the underlying structure as well.
Once it has explained the languages and concepts, the book’s description of logic and inference rules is a welcome introduction to the rationale for doing all of this. To be sure, at this
point the material gets noticably more difficult, but the description remain clear to the reader willing to dig into it. The tutorial on predicate logic was a welcome breath of fresh air for this
reader who’s more than a few years past college and his study of such things.
The final section wraps up the topic with sample applications.
Each chapter includes exercises for the student. This is appropriate for use with a college course or if you are pretending to be taking a college course.
An excellent book!
