ASP-PROLOG: Composition and Interoperation of Rules.

Abstract

One of the main goals of the Semantic Web initiative [3] is to extend the current Web technology to allow for the development of intelligent agents, which can automatically and unambiguously process the information available on millions of web pages. It has been recognized very early in the development of the Semantic Web that rules are essential for the Web3 and for Semantic Web applications—e.g., description of semantic web services, rules interchange for e-business applications. The RuleML initiative is a response to the need of a shared rule markup language using XML markup, which has a precisely defined semantics and efficient implementations. In recent years, a significant amount of work has been devoted to develop knowledge representation languages suitable for the task and a variety of languages for rule markup has been proposed. The initial design [4] included a distinction (in terms of distinct DTDs) between reaction rules and derivation rules. The first type of rules is used for the encoding of event-condition-action (ECA) rules while the second is meant for the encoding of implicational/inference rules. Despite the fact that many different proposals for ECA rules encoding have appeared the work on ECA rules is still very vague. The most recent modularized description of RuleML [6] reports this area (indicated as PR RuleML in that document) as work in progress. The derivation rules component of the RuleML initiative has originated a family of languages.4, Datalog plays the role of a core language, with simplified versions (unary and binary Datalog) developed for combining RuleML with OWL (as in SWRL). Various sublanguages have been created to include features like explicit equality (e.g., fologeq), negation as failure (e.g., naffolog), and Hilog layers (e.g., hohornlog). Various authors [7] have argued that any realistic architecture for the Semantic Web must be based on various independent but interoperable languages, including logic programming languages with and without negation-as-failure. The need for these languages and their interaction have been discussed (e.g., [8, 7]). It is also of