APIs for Rule Engines Link to API4KB Main Page

• Need to identify a target set of rule languages and rule types (derivation, reaction (production, ECA, CEP)
• Need to determine whether support for machine learning, planning, etc. is in scope long term
• Need support for a general rule language interchange (RuleML as general rule interchange language and maybe RIF as specialized Web rule interchange format (only support for basic logic rules))
  • Translate and Load vocabularies
  • Execute rules (transformations from the platform-independent RuleML interchange format into the platform-specific executable rule language)
  • Interchange of queries and answers (at a minimum; supported by RuleML)
  • Interchange of rule sets / rule bases
  • Interchange of proofs
  • Interchange of test cases for verification and validation (see verification, validation and integrity testing below)

Need to see how this connects with the work that is going on for PRR and Decision Management at OMG

• Need support for a rule interface description for distributed rule bases (see RuleML)
  • description of supported semantics
  • description of public signatures which can be queried (derivation rules) or triggered (reaction rules) including
  • description of life cycle management (e.g. descriptive metadata and qualifying metadata, e.g. validity of the rules, rule management in KB etc.)
• Need to support rule inference services (looseley-coupled or de-coupled)
  • REST services
  • wrap rule inference services into agents which communicate via messages (see e.g. Rule Responder and Messaging in RuleML)
    • support for agent role models
    • pragmatic primitives for coordination and negotiation between agents (e.g. FIPA ACL primitives)
  • description of non-functional properties of the rule service (see Rule Based Service Level Agreements)
• Need support for validation, verification and integrity (VVI) testing
  • Verification ensures the logical correctness of a rule base
  • Validation is concerned with the correctness of a rule-based system in a certain domain or application (e.g. a derived discount of 10% and 5% might be mutual exclusive or might add up depending on the domain/application)
  • Integrity (constraints) are a way to formulate consistency (or inconsistency) criteria
  • Uses Test Cases for VVI (see RuleML test cases for VVI Verification, Validation, Integrity of Rule Based Policies and Contracts in the Semantic Web)
    • Test Cases can be used for testing at design time (test-driven development) and run time (e.g. rule translation and compliance test cases)
    • Test cases can be interchanged together with rules and
      • used to validate the interchanged rule set in the target system/environment.
      • Newly received rules can be safely added to an existing rule set, if they pass all existing test cases.
      • Test cases can be predefined by experts while rules are written or combined by “normal” practitioners. The test cases then safeguard this latter authoring process.
      • Test cases are less abstract than the general purpose rules and therefore easier to comprehend for the end user.

Questions