The authors propose a formal model for the compositional verification of knowledge-based systems. The paper is well structured and organized. It uses clear and simple English expressions that are easy to understand. The formal definitions are introduced in a rigorous manner, and many other interesting definitions to support a stepwise system verification using an axiomatic base are also introduced. The association between knowledge and system is original and innovative, and very interesting to read. The compositional verification method imposes some structure on the verification process. However, the authors focus only on one level of abstraction (information and process hiding).

The model used to illustrate the compositional verification method is a generic process model for the dynamics of diagnostic reasoning. It is not developed in enough detail to fully convince readers of the utility of the formal approach proposed in the paper.

Much value would be added to the paper if the authors illustrated their work through a concrete system, using their own definitions and formulas. In addition, the quality of a decomposition and its complexity should be discussed; further clarification should be provided regarding the complexity of the compositional verification, with respect to the complexity of the system. In the bibliography section, it would be good to have some references about the system architecture optimization. Overall, the authors have produced interesting work, and should be encouraged to work further.