The proceedings of the Third International Workshop on Rules and Rule Markup Languages for the Semantic Web (RuleML), held in Japan in 2004, are contained in this book. Its main target audience is researchers working in the areas of the semantic Web, rule-based systems, and related fields. This group will benefit from reading the book; the book reflects the state of the art in this area. The book, however, is not an introduction, and is not for someone who has no knowledge of semantic Web concepts, or who is not familiar with rule-based systems.

The proceedings contain 18 papers, of which two are invited papers from keynote speakers. It starts with a paper about use cases, which motivates the need for semantic Web rules. This part is very short, and this is the weakness of the paper; it’s on a very interesting topic, but is far too short. The following paper explains the standards in this area, and includes a short example. These two papers are the ones nonexperts can also read and benefit from. The remaining papers include 11 regular papers, followed by short tool presentations.

Most papers address some extension of Rule Markup Language (RuleML), including extensions toward nonmonotic and defeasible logic, extending Web ontology language (OWL), and so on. Unfortunately, there are few papers addressing real-world examples.

Since I cannot describe all of the papers here, I will discuss two that are of particular interest to me: one “regular” paper, and one short tool presentation. Personally, I found these papers to be very interesting, as I am currently involved in a project on extracting information from tourist Web sites, and assessing the correctness of the information presented to the user. One of the more practically oriented papers (Badica and Badica, pages 37-48) presents an interesting example of feature value extraction from Hypertext Markup Language (HTML) product information sheets, and explains how rules can be generated from that. Basically, the authors convert the HTML document to Extensible HTML (XHTML), extract some examples from the set of documents, parse the XHTML document into a relational representation, and feed this to an inductive learner. As a result, they get some rules that can then be used to extract information from other documents. This paper gave me some useful ideas about how we could approach our problem using rules and RuleML. The second paper, from the short tool presentation section, comes from Jing-Ying Chen (page 182-187), and describes how rule-based services can be integrated into a Web services framework. The goal is to create a “fully dynamic service customization framework.” Rules form the glue between the different services, and a rule language is proposed for this purpose. Whether or not this approach is really useful in the Web service universe will have to be proven by practice, but at least it is an interesting idea.

As is always the case with proceedings, there is no need to start with a specific chapter; one can read the chapters in any order. For researchers working in this area, these proceedings are a must-read.