Domain Knowledge Structure, Knowledge Representation and Hypotheses Testing

Intelligent problem solving environments (IPSEs) offer students the opportunity to acquire knowledge while working on a sequence of problems chosen from the domain. Up to now we have developed several IPSEs for various curricula and applications (computer science, configuration problems, pneumatics, economic simulation games, causal modelling, and chemistry). On the surface being very different all IPSEs follow a common design theory: the student is encouraged to acquire knowledge by stating and testing hypotheses. One of these differences is the structure of domain knowledge. The aim of this paper is to show different realizations of the hypotheses testing approach for differently structured domain knowledge. The domains differ in their knowledge representation, tree-like structure vs. graph-like structure, and their possibility to absolutely locate errors. The interrelations between the domain knowledge, the representation of the diagnostic knowledge, and the contents of the help information are described. First we give a brief overview of our knowledge acquisition theory. Two IPSEs with differently structured knowledge follow. The diagnostic components and their relationships to hypothesis testing are discussed and a comparison is made that points at the commonalities and the differences between the systems that can be traced back to different domain structures.