Runtime Modelling The Novice-Expert Shift in Programming Skills on a Rule-Schema-Case Continuum

This paper describes an approach to model students' knowledge growth from novice to expert within the framework of a help system, ABSYNT, in the domain of functional programming. The help system has expert knowledge about a large solution space. This is necessary because especially novices often produce "unusual" solutions. On the other hand, it requires a model of the students' actual state of domain knowledge in order to provide user-centered help. The model distinguishes between knowledge acquisition and knowledge improvement. Knowledge acquisition is represented by augmenting the model with expert planning knowledge represented as rules. The acquisition of malrules is possible, too. Knowledge improvement is represented by rule composition. In this way, the knowledge contained in the model can be located on a gradual continuum from general rules to more specific schemas for solution fragments to specific cases (= example solutions).