James, Thanks for sharing the reference below. Do you have any other references on expert system testing? Or expert system development methodologies in general?
I have not seen a whole lot of literature on the subject, and in practice it is a major concern for us: well over 50% of our ongoing effort is in the area of testing (even with automated tools and built-in diagnostic rules). We are finding the declarative model for software systems can be powerful, but demands a very high level of verification and validation for real-world use. And many of the most robus testing methods for the procedural model simply don't fit. I've been toying with thoughts on rule and knowledge base meta-data analysis (i.e. patterns in the rules, facts and relationships between them), but would rather read about it than stare at the ceiling at night. . Hicks appears to take the approach of making the meta-data the source, and then working on this source for validation. Besides the value in testing, I think there would be great value in a consise meta-data definition of a system for the purposes of enhancment and design. I wonder if anyone has done anything to extract meta-data from CLIPS/JESS source? Anyone else out there have thoughts on this subject? Jack -------------------------------------------------------------------- To unsubscribe, send the words 'unsubscribe jess-users [EMAIL PROTECTED]' in the BODY of a message to [EMAIL PROTECTED], NOT to the list (use your own address!) List problems? Notify [EMAIL PROTECTED] --------------------------------------------------------------------
