The Semantic Web is a worldwide endeavor to advance the Web by enriching its content with semantic metainformation that can be processed by inferen- enabled Web applications. Taxonomies and rules, along with their automated reasoning techniques, are the main components of Semantic Web ontologies. Rule systems are considered to be a major area in the further development of the Semantic Web. On one hand, rules can specify declarative knowledge in ontology languages, expressing constraints or transformations, either in conju- tionwith, orasanalternativeto, descriptionlogics.Ontheotherhand, rulescan specify behavioral knowledge, enforcing policies or reacting to events/changes. Finally, rule markup languages such as RuleML allow us to publish rules on the Web, to process rules in general XML environments as well as special rule engines, to exchange rules between di?erent applications and tools via XSLT translators, as well as to embed rules into other XML content and vice versa. This workshop was dedicated to all aspects of rules and rule markup l- guages for the Semantic Web. RuleML 2004 was the third in a series of wo- shops that was initiated with the International Semantic Web Conference. The previous workshops were held on Sardinia, Italy (2002), and on Sanibel Island, USA (2003). Thisyearwehad25submissions, ofwhich11wereacceptedasregularpapers and another ?ve as short papers describing tools. Wearegratefultoourtwoinvitedspeakers, MikeDeanfromBBNandChr- tine Golbreich from the University of Rennes. Our thanks also go to all subm- ters and reviewers without whom the workshop and these proceedings could not have succe

As in other fields, in computer science certain objects of study can be synthesized from different basic elements, in different ways, and with different resulting stabilities. In subfields such as artificial intelligence, computational logic, and programming languages various relational and functional ingredients and techniques have been tried for the synthesis of declarative programs.
This text considers the notions of relations, as found in logic programming or in relational databases, and of functions, as found in functional programming or in equational languages. We study a declarative integration which is tight, because it takes place right at the level of these notions, and which is still practical, because it preserves the advantages of the widely used relational and functional languages PROLOG and LISP. The resulting relational and functional language, RELFUN, is used here for exemplifying all integration principles.

This volume presents the proceedings of an international workshop on the processing of declarative knowledge. The workshop was organized and hosted by the German Research Center for Artificial Intelligence (DFKI) in cooperation with the Association for Logic Programming (ALP) and the Gesellschaft f}r Informatik (GI). Knowledge is often represented using definite clauses, rules, constraints, functions, conceptual graphs, and related formalisms. The workshop addressed such high-level representations and their efficient implementation required for declarative knowledge bases. Many of the papers treat representation methods, mainly concept languages, and many treat implementation methods, such as transformation techniques and WAM-like abstract machines. Several papers describe implemented knowledge-processing systems. The competition between procedural and declarative paradigms was discussed in a panel session, and position statements of the panelists are included in the volume.