Expert System Ruleset Portability Using the Language Abstraction for Rule-Based Knowledge Systems (LARK) Engine

This thesis describes the Language Abstraction for Rule-based Knowledge-systems (LARK) Engine. The goal of this engine is to process various expert system rulesets and generate the required semantics for multiple production systems – thus creating true portability for expert systems such as M.1 and CLIPS. Specifically, LARK provides ruleset translation from Lark Markup Language (LarkML, an XML language defined herein), to CLIPS and M.1 expert system rules, as well as an implementation of rules written in natural language. LARK also demonstrates the ability to parse and convert basic CLIPS and M.1 rules to LarkML. In addition to describing the LARK Engine, this thesis also outlines an overview of significant expert system, UML, and business ruleset portability efforts. Ruleset portability is quickly evolving as the combined efforts of many organizations push the technology forward. Significant ruleset portability efforts include the Production Rule Representation (PRR) as defined by the Object Management Group (OMG), the Rule Interchange Format (RIF) as specified by W3C, the Rule Markup Language (RuleML) Initiative composed of a large group of industry and academia participants, and the Natural Rule Language (NRL), an effort sponsored by SourceForge.