Browse Books

Reviewer: Raphael M. Malyankar

The concept of the semantic Web is drawing an increasing amount of attention in the information technology (IT) arena. An important part of semantic Web technology is based on description logic, for example the US Department of Defense's Defense Advanced Research Projects Agency (DARPA) Agent Markup Language (DAML), and its successor, the Web ontology language OWL. The origins of description logic (DL) predate the World Wide Web, stemming from early artificial intelligence (AI) research in knowledge representation. Interest in this field has been steadily growing since the advent of the idea of adding semantics to Web information (an idea that dates back to the early days of the Web, but that has only recently gained recognition as being practical). While there are many papers relevant to the field, they are scattered among various journals and conference and workshop proceedings. Until now, it has been difficult to locate a single text that covers enough of this field to serve as a thorough introduction to it. This handbook is an attempt to fill this gap, and it does so admirably. The three main sections of the handbook concern theory, implementation, and applications. An introductory chapter traces the development of the field, sketches its basic concepts, surveys application domains, and outlines the connections between DL and other areas of computer science and AI. An appendix containing definitions of DL terms is also included. The five chapters in the theory section of the book (Part 1) offer a thorough introduction to the established theory of description logic. Chapter 2, "Basic Description Logics," is an excellent tutorial on basic theoretical ideas, terminology, and inferencing methods. Chapter 3, "Complexity of Reasoning," addresses exactly that; it consists of an analysis of the computational complexity of subsumption and satisfiability. Chapter 4, "Relationships with Other Formalisms," is a survey of the commonalities and differences between DL and frame systems, conceptual graphs, entity-relationship (E-R) models, object-oriented models, Extensible Markup Language (XML), and a selected set of other (non-DL) logic-based formalisms. Chapters 5 ("Expressive Description Logics") and 6 ("Extensions to Description Logics") extend the exposition of chapter 2, by introducing more advanced theoretical ideas. Overall, this first section is thorough, and should be able to stand by itself as a useful learning resource, though space constraints and the nature of the material have no doubt dictated the omission of detail in places, and the student will have to refer to the cited sources for detailed expositions of some of the results mentioned. Part 2, "Implementation" (chapters 7 through 9), is more mixed in its focus and depth of treatment. Chapter 7, "From Description Logic Provers to Knowledge Representation Systems," is a good general discussion of what is necessary in constructing any deployed knowledge-based system. Chapter 8, "Description Logic Systems," surveys some implemented systems (both DL systems and certain precursor knowledge representation systems). Chapter 9, "Implementation and Optimization Techniques," complements chapter 3, with a disquisition on the issues encountered in implementing reasoning algorithms in DL systems, providing an informative round-off to the largely theoretical content of Part 1. "Applications," the third part of the book, consists of seven chapters on different application areas, with mixed foci and treatments. Chapter 10, "Conceptual Modeling with Description Logics," is a short tutorial on information modeling considerations, with special attention paid to the DL issues involved. Chapters 11 and 12 briefly discuss DL applications in software engineering and configuration, mentioning the representation and reasoning issues that arose in systems implemented in these domains, and how description logic was used or could be used to address these issues. Chapter 13, "Medical Informatics," is a fairly detailed overview of medical informatics modeling and of the challenges it presents for DL representations of medical information, including brief discussions of specialized features used for different problems in the representation of medical knowledge with DL. Chapter 14, "Digital Libraries and Web-Based Information Systems," is largely a description of DAML, ontology inference layer (OIL), and DAML+OIL, precursors of the OWL Web ontology language. It also summarizes early examples of applying semantics to Web information (specifically, Untangle and FindUR). The short chapter on natural language processing (chapter 15) describes efforts in using DL representations in natural language processing research (acknowledging that advances in other techniques applied to natural language processing problems have overshadowed the use of DLs in this area). Chapter 16, "Description Logics for Databases," briefly summarizes the application of DL to data models (particularly the E-R model), database querying, and data integration. Some chapters in this part, namely chapters 14 and 16, may be better classified as derived technology or related research rather than domain applications, which is what an implementer might be expecting to see. This part will be particularly useful as a collection of experience reports, with hints on how to address specific problems that may arise while implementing a DL-based system. More details, and more comprehensive indexing and cross-referencing of the material in this section, would make such use of it easier. One may categorize the material in the book as follows: chapters 1 through 3, 5, 6, 9, and the appendix together describe the nuts and bolts of description logic; chapters 4, 10, 14, and 16 focus on derived or related formalisms, technology, and theoretical constructs; chapters 11 through 13, 15, and part of chapter 16 describe applications in different domains; and chapters 7 and 8 focus on systems deployment considerations and history. A useful addition to the next edition would be one or more in-depth case studies of actual DL systems: there are several systems discussed in multiple chapters of the book, but a really detailed case study would, I think, greatly enhance the reader's understanding of the subject. This book is suitable for both academics and implementers, though the latter might prefer to skim much of the theoretical exposition in Part 1. Prior knowledge of logic, or exposure to knowledge representation systems or formalisms, will help in understanding the book (particularly the theory section); readers without such exposure will find the book more difficult. The book would serve well as a foundation text for a graduate course in description logic, when supplemented with selected research papers. The production qualities are excellent; this book will be easy to read and refer to. Description logic and other areas in semantic Web technology are currently seeing some very active research, so some of the contents of the book will be dated and the book itself will be incomplete in a few years' time. This cannot be helped, however. The book, as it stands now, is a needed and well-executed reference work. Some more material about advanced topics would have been nice to see, as well as chapters on what cannot be done with description logic, and on open research areas in both DL itself (mentioned briefly in some chapters) and in related fields such as Web services. These are only minor quibbles with regard to an excellent book. Online Computing Reviews Service