The study of social networks has been around for a while now, but the application of graph theory and computer methods has given a more analytic face to the topic. This promises to be a fascinating area of research in the future, as these techniques are used to analyze everything from how communities of people are built to how organizations are run.

In particular, the Internet and Web provide sources of data that can be used for the analysis of social networks. Consider all the possibilities: email, chat systems, online games, and links from Web pages. They all reflect in some way the social interactions of people, and the emergence, growth, and sometimes decline of communities. And, of course, there are the specifically social Web sites, such as MySpace and Facebook.

The semantic Web is an emerging set of technologies intended to provide automatic (machine) interpretation of information and derive meaning from that. The semantic Web is often built on ontologies--descriptions of something (or of related things) using an ontology representation language. Ontologies are usually built of triples, essentially subject-connector-object groupings. For example, a simple triple, in no specific ontology markup, might be “Saturn-rdf:type-planet,” meaning that Saturn is a planet. Since ontologies are intended to be comprehensive, the term planet itself is likely to be described somewhere in an ontology, to which the ontology describing Saturn may link; “rdf:type” is a defined connector in the resource description framework (RDF) ontology description notation.

Semantic Web technologies seem a natural fit for studying social networks. Since the ontologies are both descriptive and flexible, they can be used to describe new kinds of relationships as they are discovered, and their description serves both to define the new kind of relationship and provide a way for algorithms to explore it.

This book is about the intersection of semantic Web technologies and social networks. It provides an overview of both topics, and detailed information about how ontologies are represented and how social network data is modeled. It then provides three chapters on extracting social network information from the Web; on modeling social networks in scientific communities, particularly in the community of scientists focused on the semantic Web; and on folksonomies. It finishes with a chapter covering the emergence of Web-based systems for locating people after hurricane Katrina.

The strongest parts of the book are its overviews of social networks and the semantic Web. The sections on social networking are very good indeed, and might well serve to draw people from divergent disciplines into a fascinating field of research. The overview of the semantic Web and of the various ontology representation languages is also good, but the subject is very complex and difficult, and the reader will need to look elsewhere for enough coverage to answer anything but the simplest questions.

Later sections are weaker. The chapters seem rushed, and are not well edited. For example, in chapter 7, on Web-based social network analysis, Figure 7.3 is almost unreadable, and the x and y axes have no meaningful explanations. Figure 7.1 shows that one individual seems to have produced more than a million Web pages--this is a remarkable fact, as it would mean that one person has produced 160 Web pages a day, every day, since 1990. Other figures are many pages away from their references (in one case, ten pages), or are hard to read. One figure purports to show DARPA Agent Markup Language for Services (DAML-S) researchers forming a cluster away from the core researcher in the area, but how it shows that is beyond my ken.

The section on social networks in the semantic Web community seems to be more of a discussion of that community than a serious analysis of the social networks involved.

In a few places, Mika is crusading against Extensible Markup Language (XML) in favor of RDF or other ontology representations. There is much to be said about XML’s limitations and problems--and many who say it--but RDF, XML, and other forms of data markup, such as the tuples in a relational database, do not necessarily compete to solve the same problems. In particular, in the discussion on the people finder system developed in the aftermath of hurricane Katrina, the inflexibility of a quickly developed XML schema is pointed out as a weakness of the representation. It is far from clear how the flexibility of RDF, and the tools available to use it at that point, would have led to a better result. While this is likely to change with the emergence of new tools, many of which are mentioned in the book, it seems, at this point, that each of the various technologies has specific strengths and weaknesses, and no one will entirely replace the others. The focus should be on helping these technologies work together.

One of the really strong points of the book is the collection of Web links provided in the footnotes. Following just a few of these links has already proven their worth, and it is likely that the rest will be just as interesting.