The article discusses how relational-product architectures can unify different types of information processing methods, such as classical, probabilistic, and fuzzy inference. These architectures are useful for creating intelligent knowledge-based and expert systems. The authors show that relational products can represent various types of relationships in a more flexible way than traditional database designs. This approach offers a wider range of structures and algorithms for manipulating data.