In this dissertation, we present a generic methodology, GAAOP, for adding adaptivity to educational games. A detailed overview of the multi-disciplinary domain of adaptive educational games is given, in order that readers with background in either the Computer Science, Cognitive Psychology, or Engineering can understand the framework. This includes definitions and interpretations of adaptivity for the domain of Cognitive Psychology and for the domain of virtual reality, patterns and architectures used in the domain of game development, and Aspect-Oriented programming techniques that target the issues related to programming cross-cutting concerns.
The methodology was created via an inductive approach, where a general method was extracted from the manual implementation of adaptivity to an example educational game, TuxMath, in the context of a pilot study for the CAdE project, a multi-disciplinary research project on cognitive adaptivity for educational games. Where the first implementation required the programmer to introduce adaptive behaviour in between the original game code, GAAOP provides a framework of aspects that allows the programmer to separate the adaptivity concern from the rest of the game. As a proof of concept, this general pattern was applied to a simple game that can serve as a test case for implementations in different languages and aspect systems.
The resulting framework is generic enough to be applied to any game, and can be extended with more specific reusable behaviour to lower the implementation effort, and by extend also the development cost.
Compared to other solutions, GAAOP stands out in separating the adaptivity concern from the original game code. The focus of other solutions lies more with interpretations of adaptivity, indicating that our work is complementary with existing work.