In this paper, we present an overview of TUBERDIAG, a rule-based expert sys
tem for diagnosis of pulmonary tuberculosis. This system was developed as a
2-year joint project by the Institute of Information Technology (National
Center for Science and Technology) and the National Institute of Tuberculos
is and Lung Diseases, Hanoi. After designing and building a suitable infere
nce engine for this system, we have done a lot of work to create an effecti
ve knowledge base; at present, this knowledge base contains more than 1,000
rules. The paper focuses on how the rule base is constructed, managed, and
used. We also present here the first evaluation of TUBERDIAG by the doctor
s at the National Institute of Tuberculosis and Lung Diseases, who have bee
n playing a very important role in the project - providing the system with
their knowledge.