The project will assess several design alternatives in the development of expert systems and, more generally, machine intelligence. Specifically, a universal knowledge frame will be evaluated as a means of: 1) reducing the ad hoc nature of the knowledge base of expert systems, 2) facilitating the production of new expert systems, 3) permitting the interaction of expert systems from within their distinctive domains, and 4) allowing for more robust semantic processing capability. A limited protype of the universal knowledge frame will be implemented through a relational database (extended to support semantic data modeling) coupled with rule-based control mechanisms defined in prolog. Each of the implementational techniques will be assessed, and the terms of practical software development potential.