Date: Wed, 29 Sep 2004 19:09:00 +0200 (MEST) From: Scott Farrar Subject: Spinning the Semantic Web
EDITORS: Fensel, Dieter; Hendler, James; Lieberman, Henry; Wahlster, Wolfgang TITLE: Spinning the Semantic Web SUBTITLE: Bringing the World Wide Web to Its Full Potential PUBLISHER: The MIT Press YEAR: 2003
Scott Farrar, University of Bremen
SUMMARY OF CONTENT This collection is one of the first major works covering the origins, state of the art, and enabling technologies of Tim Berners-Lee's vision of the Semantic Web. With 41 contributors, many of whom major players in the emerging Semantic Web, this collection offers a broad view of the Semantic Web by including contributions ranging from the theoretical to the practical.
Foreword by Tim Berners-Lee. Based on a talk given by Berners-Lee in 1997 at the London W3C meeting, the foreword provides philosophical and practical motivation for the Semantic Web (SW) and the rest of the book. Berners-Lee discusses the current state of the Web and argues for a Web based metadata, trust, and, in general, the technology being developed by the World Wide Web Consortium. In particular, the author stresses the need for the Resource Description Framework (RDF).
Chapter 1. 'Introduction', by Dieter Fensel, James Hendler, Henry Lieberman, and Wolfgang Wahlster. If the foreword precedes from futuristic idealism, the point of view of the introduction is one of practicality for the here and now. That is, the editors of the volume motivate the SW by highlighting the benefits to knowledge management, Web commerce, and electronic business. The various enabling technologies are introduced in the context of the SW: the Extensible Markup Language (XML), Extensible Hypertext Markup Language (XHTML), RDF, RDF Schema (RDFS), etc. Tools for working with such languages are introduced, including Protege-2000. Applications are adduced including On-To-Knowledge.
Part I: Languages and Ontologies Chapter 2.'SHOE: A Blueprint for the Semantic Web', by Jeff Heflin, James Hendler, and Sean Luke. Motivated because of the distributed, dynamic, massive, and open-world aspects of the current Web, an ontology language called SHOE is introduced and described. After a detailed background to the approach, the SHOE language is formally defined and an example SHOE ontology is given in XML. Issues of interoperability and evolution of ontologies are presented, followed by a description of an implementation from two test domains. Related works are also discussed.
Chapter 3.'DAML-ONT: An Ontology Language for the Semantic Web', by Deborah L. McGuinness, Richard Fikes, Lynn Andrea Stein, and James Hendler. The authors present the history and motivation for the DAML ontology language, DAML-ONT. The authors discuss the Web language through a series of examples in XML. Next, they present the axiomatic semantics of DAML-ONT in a rigorous format, again with examples in XML, but also including examples in RDF and the Knowledge Interchange Format (KIF).
Chapter 4.'Ontologies and Schema Languages on the Web', by Michel Klein, Jeen Broekstra, Dieter Fensel, Frank van Harmelen, and Ian Horrocks. In this chapter the authors discuss the relationship between ontologies and the various schema languages of the Web (XML Schema and RDF Schema), namely for the purpose of investigating how these schema languages can be used to add semantics to Web content. The authors give an abstract discussion of ontologies and schema languages and their relationships. Next, an overview of the Ontology Inference Layer (OIL) language is given. They discuss elements of the OIL language and give various examples. Next, an overview of XML Schema is given. The authors then compare OIL and XML Schema by looking at the different uses the languages were intended for. Next, an overview of RDF Schema with many examples is given, followed by a comparison of OIL and RDF Schema. An in-depth example of applying ontologies to online resources in XML is presented in a step-wise fashion. Particular attention is given to how to create hierarchies and type constraints.
Chapter 5.'UPML: The Language and Tool Support for Making the Semantic Web Alive', by Borys Omelayenko, Monica Crubezy, Dieter Fensel, Richard Benjamins, Bob Wielinga, Enrico Motta, Mark Musen, and Ying Ding. The aim of this chapter is to present the Unified Problem-Solving Method Development Language (UPML) framework and associated tools and to demonstrate how this language provides a key enabling technology for the SW. The UPML is presented as a framework to mark up reusable knowledge resources and components. The authors present and detail the IBROW system as an example of an application that takes advantage of UPML marked up resources. Then, an in-depth description of the UPML itself is presented with many examples in various Web syntaxes.
Chapter 6.'Ontologies Come of Age', by Deborah L. McGuinness. The aim of this chapter is to motivate the use of ontologies as laid down by Tim Berners-Lee (see the Foreword). The author presents an introduction to ontologies and provides several simple examples. The author then contrasts the use of simple ontologies with more structured ones which offer consistency checking, completion, and interoperability. Then, the notion of ontology acquisition is discussed. Finally, the author gives the various enabling technologies for ontologies, including languages (KIF, DAML+OIL, and XML) and development environments that are collaborative scaleable, and interconnected.
Part II: Knowledge Support Chapter 7.'Sesame: An Architecture for Storing and Querying RDF Data and Schema Information', by Jeen Broekstra, Arjohn Kampman, and Frank van Harmelen. The authors introduce and describe the Sesame architecture and motivate its development from the need to develop an efficient way to store SW data. As background, they then introduce and detail RDF and RDF Schema languages. They argue for a query language for RDF(S) and introduce the RDF(S) Query Language (RQL). The authors then turn to the Sesame architecture and discuss the various components, including the repository, repository abstraction layer and the PostgreSQL language. The functional modules are presented and experience using the system is given.
Chapter 8.'Enabling Task-Centered Knowledge Support through Semantic Markup', by Rob Jasper and Mike Uschold. The authors' aim is to motivate the need for various SW components that locate resources, appeal to the users' needs, and provide a semantics. Various facets of Web problem solving are discussed. The authors then turn to the concrete domain of aerospace customer service and present several examples of problem solving. The enabling infrastructure for such a Web is given with an emphasis on semantic metadata and information and knowledge repositories. They conclude with a worked example in XML with semantically oriented tags.
Chapter 9.'Knowledge Mobility: Semantics for the Web as a White Knight for Knowledge-Based Systems', by Yolanda Gil. Arguing for more knowledge 'mobility', the author introduces the TRELLIS system which aids in the tracking of data/knowledge so that knowledge will not be confined to a single formalism. First, the author gives a very detailed introduction to the notion of mobility and provides examples such as translation between agents. The author then gives a short description of Resilient Hyper-Knowledge Bases (RHKB) and the TRELLIS system that builds them.
Chapter 10.'Complex Relationships for the Semantic Web', by Sanjeev Thacker, Amit Sheth, and Shuchi Patel. The primary topic of this chapter is the nature of complex relationships on the SW with an eye toward establishing a deeper for semantics for Web content. As a practical approach, the authors describe the InfoQuilt which is provides a framework and a platform for answering complex information requests. They next give many examples of complex relationships. This motivates the need for more complex queries, or 'requests'. This notion is then expanded to the 'information scape', or IScape. The authors argue that IScapes enable better knowledge discovery and motivate their stance with a series of examples. The need for better visual interfaces for IScape usage is presented and adduced by a series of examples. Finally their approach is contrasted in a section on related work.
Chapter 11.'SEmantic portAL: The SEAL Approach', by Alexander Maedche, Steffen Staab, Nenad Stojanovic, Rudi Studer, and York Sure. This chapter deals with how Semantic Web data should be communicated between the provider and the consumer of the data. The goal is to institute a methodology, which the authors call the 'SEAL' approach. First, the authors overview the notion of ontology and in particular ontologies for communication. They contrast ontologies with knowledge bases and give a formalized definition of both. Then, a methodology is given for the knowledge engineering process at the kickoff, refinement, and evaluation stages. The authors then describe the SEAL architecture in depth including a description of each of the modules. In the next section a process called 'semantic ranking' is introduced and it is detailed how this notion relates to queries and the comparison of knowledge bases. Furthermore, the notion of semantic personalization is introduced as part of the SEAL architecture. The last part of the chapter discusses RDF and how the tool RDFMaker can be used to transform data in a data base system to RDF. Finally, the authors discuss related work in connection with each component of SEAL.
Part III: Dynamic Aspect Chapter 12.'Semantic Gadgets: Ubiquitous Computing Meets the Semantic Web', by Ora Lassila and Mark Adler. Because the discovery of semantic services is seen as a key aspect of the SW in the context of ubiquitous computing, the authors of this chapter discuss how so-called 'semantic gadgets', or devices/applications that utilize the SW, can be guided to discover and use such services. First, the authors give a brief overview of knowledge representation and provide an example in the museum domain. Key notions of knowledge discover are overviewed and briefly discussed. The notions of 'contracting for use' and 'composition of services' are presented, following by some analysis of the museum domain.
Chapter 14.'Semantic Annotation for Web Content Adaption', by Masahiro Hori. The aim of this chapter is to present a methodology whereby SW content can be 'transcoded' for small devices. The author begins with an example of how an external annotation framework that uses XPATH can be used to annotate HTML. The author then gives an annotation-based transcoding system. One example given is how an HTML page can be split for small-screen devices. An example of a real-life webpage is worked through with the annotation framework. In the final section, the discussion includes various empirical results obtained in the transcoding process.
Chapter 15.'Task-Achieving Agents on the World Wide Web', by Austin Tate, Jeff Dalton, John Levine, and Alex Nixon. This final chapter focuses on the Web as a place where one can 'do' things. The topics include: ontologies for plans and processes in the military domain and collaborative multiagent-human systems. First the authors give an overview of the standards for representing activities. The ontology for activities is then presented and motivations are given. Next, the authors describe the generic interface for Web-based, task-achieving agents called Open Planning Process Panels (O-P^3). Finally, three Web-based applications are described: the O-Plan Web demonstration, the Air Campaign Planning Process Panel and a wireless version of O-Plan.
CRITICAL ANALYSIS The primary merits of the reviewed collection include its extremely broad scope. The content of the major topics---languages and ontologies, knowledge support, and dynamic aspect---is well represented by diverse contributions that vary according to depth and approach. Another major merit includes the ''all-star'' author line-up which gives the collection a certain respectability in the emerging field of the Semantic Web. At the same time this work can be criticized for attempting too much too soon. With the vision of the Semantic Web largely still largely unclear, its tone slightly too optimistic, with some contributions ignoring issues such as the broad acceptance of the Semantic Web and a migration methodology. Another problem is that many of the contributions overlap. For example, there are various introductions to ontologies, and the coverage of RDF is repeated more than once. This collection is no Semantic Web cookbook for the beginner. Whereas some chapters are introductory in nature, no generalized, unified approach is given. This is not a criticism of the book, just a caveat for the potential reader. In general this collection provides a well-structured introduction to the field, covering the basics but also moving beyond the fundamentals.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER Scott Farrar is currently a post-doctoral researcher at the University of Bremen in Bremen, Germany as part of the Spatial Cognition 'Sonderforschungsbereich' sponsored by the German Government. His chief current research interests include the Semantic Web, linguistic knowledge bases, ontologies and spatial cognition; see his website at: http://www.sfbtr8.uni-bremen.de/i1