Within the framework of so-called second generation expert systems
[62] knowledge modeling is one of the most important aspects. On the
one hand, knowledge acquisition is no longer seen as a knowledge
transfer process, rather it is now considered as model construction
process which is typically a cyclic and error prone process. On the
other hand, the distinction between knowledge and symbol level
descriptions [166] resulted in various proposals for conceptual
knowledge models describing knowledge in an implementation independent
way. One of the most prominent examples of such a conceptual model is
the KADS model of expertise which is characterized by its clear
distinction of different know ledge types and by the usage of specific
modeling primitives to describe these different knowledge types [185].
The semi- formal KADS expertise model entails all the advantages and
disadvantages which have been identified for semi-formal system models
e.g. in the software engineering community.