Modeling Domain Knowledge Ushg Explicit Coneepfualiza#ion

ized by the tasks and domains involved. For example, we can talk about a planning ap-plication for an elevator system, or the task of diagnosing digital devices in general. Knowledge modeling can therefore be di-vided into two conceptual subactivities: mod-eling the task and modeling the domain knowledge. If properly designed, the sepa-rate parts can be reused: Considerable parts of a task model can use different domain models, and vice versa. However, this does not always happen. Commonly, domain knowledge is reentered for every new knowl-edge-based system, making system con-struction expensive. Parts of the domain knowledge partake in many tasks and can be reused if properly specified and organized. Moreover, this or-ganization can be designed to support flexi-ble reasoning. In this article we discuss an explicit conceptualization of the domain knowledge, at the heart of its organization. A conceptualization is the objects presumed to exist and the relationships and functions among them. ’ We augment the conceptual-ization with reusable annotations to form the-ories about the domain knowledge. The an-notations and the conceptualization guide the construction of applications and support flex-ible reasoning during problem solving