Addresses the use probability theory as a tool for designing with and implementing uncertainity reasoning. Provides many concrete algorithms, explores techniques for solving multimembership classification problems not based directly on causal networks, and offers practical recommendations, matching specific methods with sample expert systems.