From: Harold Boley (boley@informatik.uni-kl.de)
Date: 02/18/03
Dear JC Colleagues, please find enclosed the current Outline version of the RuleML Working Note Draft that Said, Benjamin, and I have been working on in conjunction with the RuleML Steering Committee. Best, Harold RuleML - Semantic Web Rule Markup Language DISCLAIMER The information presented in this Table of Contents Draft is intended for the sole use of the DAML Joint Committee. This is only a preliminary Draft. Expect updated versions of the document to be accessible from http://www.ruleml.org/ in the future. Working Note Draft, 16 February 2003 Authors: Harold Boley (harold.boley@nrc.gc.ca) Benjamin Grosof (bgrosof@mit.edu) Said Tabet (stabet@attbi.com) Copyright 2000-2003, The RuleML Initiative. Abstract This document describes RuleML, the Rule Markup Language. RuleML is a markup language for publishing and sharing rule bases on the World Wide Web. RuleML builds a hierarchy of rule sublanguages upon XML, RDF, XSLT, OWL, and other W3C languages. Table of Contents Abstract and Brief Summary (BG-ST-HB) Goal of interoperable rules over/on the Web to support integration of knowledge bases comprised of rules and facts Design approach based on declarative logic programs Webized to use URI's, namespaces, and XML Encode rulebases in XML or RDF syntax Various expressive extensions done; more possible, some already in progress Synergizes with other aspects of Web standards, especially Semantic Web (RDF, OWL) and Web Services (Web Choreography, DAML-S); anticipated with Semantic Web Services in near future -- to be coordinated with Joint Committee Background (ST-BG-HB) The need for Rules on the Web Earlier work on interoperable Web rules using XML markup The Rule Markup Initiative Rules in the Semantic Web Rules in Web Services Later: Rules in other W3C efforts, e.g., P3P/APPEL, XQuery, XSLT; SWSC) Why standardize rules now RuleML Goals, Requirements, and Design Approach -- to be shortened (BG-ST-HB) Goal of interoperable rules over/on the Web to support integration of knowledge bases comprised of rules and facts Based on underlying KR with well-understood semantics, interoperation with semantic equivalence Initially targeted families of rule systems are currently commercially the most important (later: FOL theorem-prover fragments) Relational DB's, SQL-99 ISO Prolog Production rule systems (from OPS5 to Jess) ECA rules OCL (constraint part of UML) Requirements for both forward and backward inferencing Requirements for negation: Negation as failure (NAF), strong negation (NEG), and - most declaratively - NAF-NEG-neutral negation (NOT) Requirements for prioritized conflict handling Design approach: based on declarative logic programs start with Datalog as kernel, then add extensions guided by research/development maturity, practical importance, strategic/evolutionary considerations General objectives Be in spirit of the Web, accomodate its openness, exploit its other aspects: esp. URI's, namespaces, XML as syntactic vehicle Design approach of Webizing: name relations and functions, rules, rule subsets/modules -- by using URI's and namespaces Bridge XML and RDF data models, be able to exploit RDF and harmonize with it (this might help speed RDF in rules world) Tasks: inferencing, (XSLT) translation, static analysis Applications: business policies, business process automation and workflow, e-contracting, security/authorization/trust Aim to support RDF, and also to support XML directly for those not (directly) dealing with RDF Design approach: abstract (graph-theoretical) data model with labeled arcs and minimal ordered arcs; inspired by software/data engineering practice, and by RDF, alternative syntactic encodings in XML and RDF; in-progress also in OWL -- to be coordinated with Joint Committee Pursue: multiple KR expressiveness classes, which form a partial order Design approach: hierarchy of expressive classes, syntax specification (incl. DTDs and later XML Schema) for each Some defined already, others in progress, even more ones provided from RuleML community Need syntactic mechanics for this: e.g., in DTDs and XML Schema In progress: it's useful to think in terms of a few different rules types: not just derivation and transformation, but also: action/reaction; however, these not yet well worked out wrt consensus way to view them, nor in some ways wrt the fundamental research understanding Language Structure (HB-ST-BG) Rule Types: Derivation, Transformation, and (in progress) Reaction Webized Logic: URIrefs for constants, relations, and functions The Hierarchy of expressive classes Example: diamond formed by Datalog, Horn, Datalog+NAF, Horn+NAF Overview of some others: UR versions + RDF versions Plugging in Different Sets of Relational and Functional Built-ins (ISO Prolog, CommonRules, CommonLogic/KIF, Jess, etc.) The Kernel Datalog Sublanguage (HB-BG-ST) Datalog Facts Datalog Rules The Full Horn Sublanguage (HB) Complex Terms Later: List Terms as a special case The Transformational Sublanguage (HB) Function calls via nano elements Call-by-value nestings Function definitions through trans elements Negation as Failure (NAF), Normal Logic Programs -- to be udated (BG-HB-ST) English explanation: What it does, relationships to NEG and NOT Explain the need: For W3C, IR, rules, and database communities Explain the restriction: Failure Negation appearing only on selected body atoms Allowed uses of NAF and conjunctions in the body The Prioritized Sublanguage (BG-ST-HB) Rule labels Partial order for rule prioritization Later: Certainty factors and their transformation into a partial order The Object-Centered Sublanguage (HB-ST-BG) Concept; a metarole for user-defined role names; relationship to OO Relationships to positionalized sublanguages Modeling RDF resources and OWL instances -- to be coordinated with Joint Committee Later: Modeling instances of F-Logic and Case-based reasoning Later: Types and inheritance as in OWL classes and Description LP -- to be coordinated with Joint Committee Lloyd-Topor Logic Programs: More First-Order Logic Features (BG-ST-HB) Lloyd-Topor Transformation without, then with NAF Disjunction and conjunction in body Conjunction in head Later: Existentials in body Later: Universals in head Later: Universals in body Later: Negation Rulebases, Queries, and Turnstiles (HB-BG) The rulebase element for expressing and transmitting assertions The query element for expressing and transmitting queries The turnstile element for posing a query to a rulebase Relationships to DAML Query, RDF Query, and XML Query Local Modules and Global Inclusions (ST-BG-HB) Modules: Using turnstile elements in rule bodies Inclusions: Loading external rulebases Reaction Rules (cf. "V1 Draft" of Reaction Rules Subgroup) Notion of Event Notion of Condition (from Derivation Rules) Notion of Action Notion of Postcondition (for declarative state change) Event-Condition-Action-Postcondition (ECAP) Rules Reference Implementations (HB-BG-ST) Concepts: translation, querying and inferencing, authoring, application XSLT Translators Translators between sublanguages Translators to other languages Translators to (X)HTML for customized human-oriented renderings Query and Inference Engines Graphical Editors RuleML-based Applications Use Cases / RuleML-Based Applications (ST-BG-HB) Another document in progress, i.e., from Joint Committee Animals Rulebase GEDCOM RACSA Supply Chain e-Contracting Rental car business rules Book Pricing Authorization Appendices (BG-ST-HB) Appendix A. The RDF-XML-Integrating Data Model for the Abstract Syntax: Types and Roles (HB) Appendix B. The Hierarchy of DTDs with Specs for all Sublanguages (HB-ST-BG) Abstract Syntax Associated Semantics Complete DTDs / XML Schemas RDF Syntax Later: N3 Syntax Appendix C. The Standard Built-Ins Based on XML Schema, Part II, Datatypes (ST-BG-HB) Appendix D. Notes (HB-BG-ST) Appendix E. Index of All Language Elements (HB-BG-ST)
This archive was generated by hypermail 2.1.4 : 02/18/03 EST