Human reasoning about non-trivial real world problems is a complex activity that frequently incorporates multiple reasoning strategies, concerning multiple domain knowledge which may be perceived as having different structures. The paper proposes the use of the "smart object paradigm" for the design of complex knowledge based systems. The paradigm provides mechanisms for both hybrid knowledge representation and multiple inference strategies. Central to the paradigm is the concept of "smart objects", engineered artifacts that combine a high level object structure with a rule based lower level language. The concept of smart object was developed using criteria that evolved in the course of development of a large, complex knowledge based system. We present these criteria as desirable characteristics for KRs in general, and use them to evaluate traditional KRs and smart objects. An overview of a prototype KBS implemented using the smart object paradigm makes its benefits concrete.