This research investigates the use of meta-interpreters for building expert systems in Prolog. An expert system is conceived as a knowledge base and a collection of simple meta-interpreters. Two main issues will are addressed: how to 'mix in' a meta-interpreter into a Prolog program, so that the program is imbued with the functionality of the meta-interpreter, and how to combine the effects of several meta-interpreters. Building on earlier work, a classification of meta-interpreters will be developed to facilitate combination. Expert systems would then be developed by building the knowledge base, describing extra features with meta-interpreters, combining them to produce a new metainterpreter, and then mixing the new meta-interpreter into the knowledge base. The necessary tools and techniques for facilitating such system development will be investigated and implemented. The importance of this work is its contribution to building expert systems with extensive knowledge bases which are still humanly manageable and interpretable.
