Abstract

We propose a software architecture for user-cooperative systems, like knowledge systems. This architecture incorporates important notions for building up a context, such as cooperation, explanations and incremental knowledge acquisition. Our main concern in this paper is representation methodology. We show that the architecture can be realized as a pure first order theory, which we argue constitutes a good basis for system development. A declarative relational language promotes modularity, clarity and maintainability. The system we outline is implemented in logic programming.

Citations