A one-week seminar on 'Rule Markup Techniques' will be hosted by the Dagstuhl International Conference and Research Center for Computer Science (Wadern, Germany) on February 3-8, 2002. "Rule systems (e.g., extended Horn logics) suitable for the Web, their (XML and RDF) syntax, semantics, tractability/efficiency, and transformation/compilation will be explored. Both derivation rules (which may be evaluated bottom-up as in deductive databases, top-down as in logic programming, or by tabled resolution as in XSB) and reaction rules (also called 'active' or 'event-condition-action' rules), as well as any combinations, will be considered. This 'Rule Markup Techniques' seminar aims at bringing together the classical- and Web-rule communities to cross-fertilize between their foundations, methods, and applications. The long-term goal is a Web-based standard for rules that makes use of, and is also useful to, the classical rule perspective. The seminar is expected to contribute to some open issues of recent proposals such as Notation 3 (N3), DAML-Rules, and the Rule Markup Language (RuleML). Furthermore, by studying issues of combining rules and taxonomies via sorted logics, description logics, or frame systems, the Seminar will also discuss the US-European proposal DAML+OIL. Two particular issues that will be addressed during this seminar are efficient implementation techniques (e.g., via Java-based rule engines) and major exchange applications (e.g., using e-business rules)." Conference organizers include internationally-recognized authorities on rule and agent markup languages: Harold Boley (DFKI Kaiserslautern, Germany), Benjamin Grosof (MIT Sloan School of Management, USA), Said Tabet (Nisus, USA), and Gerd Wagner (Eindhoven University of Technology, The Netherlands).

From the seminar description:

Ontologies for the Semantic Web can use rules to define axioms operating on a taxonomy, but also to compute the taxonomy from given concept definitions. Further purposes of rule markup include:

• Providing a rule interchange format for exchanging rules between different tools/systems
• Marking up rule content in business documents (for example, discount rules can be included in an internal XML-based product catalog distributed to dealers)
• Providing a high-level specification language for active content (such as agents) in the Web
• Expressing an interaction protocol (i.e., the public part of an inter-organizational business process) as a set of reaction rules and marking it up for communication over the Web

To accomodate the various user communities from knowledge-based systems, to intelligent agents, to e-commerce, possibilities for a modular rule-language hierarchy will be discussed. It should permit the mixing and matching of rules of appropriate expressive power as well as computational efficiency for classes of applications. This may be facilitated by a separation of concerns [between semantics and syntax]...

[WRT] Semantics: A classical hierarchy version can be bottomed on ground facts over individual constants; a Webized version could also permit URIs as arguments and names of (binary) relations, much like in RDF triples. On this basis, the classical Datalog, Horn-rule, and equational-logic layers could similarly be extended by URIs as relation, constructor, and function symbols. More advanced rule extensions could include suitable negations for the open world of the Web, feature or F-Logic terms for RDF's anonymous resources, labelings for prioritized business rules, and local rule packages for N3's contexts. The power-efficiency tradeoff should inform us whether the hierarchies, classical and Webized, can be topped by rules in full first-order or even restricted higher-order logic...

Related: the Call for Papers and Participation in connection with "Business Agents and the Semantic Web (BASeWEB)." This AI-2002 Workshop on Business Agents and the Semantic Web will be held May 26, 2002, Calgary, Alberta, Canada, in conjunction with AI 2002 (AI-2002). The Deadline for Submissions/Requests for Participation is February 28, 2002. "This workshop addresses researchers extending Web techniques by AI or transferring AI techniques to the Web in an attempt to create intelligent business agents... E-business increasingly uses Web Services or agents acting on behalf of human buyers and sellers. Such Business Agents can profit from the machine-interpretable product and service descriptions provided by the Semantic Web. Cross-fertilized techniques from AI (e.g., Intelligent Agents) and the Internet (e.g., the Semantic Web) are thus explored by numerous organizations world-wide, including W3C, DARPA, NRC, IST, and INTAP. Web ontologies - consisting of taxonomies and/or rules - constitute the centerpiece of the new AI-Internet synthesis. Topics of interest include, but are not limited to product and service codes/registries (e.g., UNSPSC/UDDI); description logics and Web ontologies (e.g., DAML+OIL, WOL); extended Horn logics and rule markup techniques (e.g., RuleML); application areas for rules (e.g., P3P, XACML, ebXML, DRM, etc.)..."