THE PRIMEGAME: COMBINING SKILLS IN UNDERGRADUATE COMPUTER SCIENCE PROGRAMMES

Foundational software engineering skills such as numerical methods, algorithm design and programming are usually taught as disjunct courses in most undergraduate computer science programmes. While those and other skills have to be combined in order to develop suitable solutions, such integration usually only takes place in advanced courses where larger projects are then often subject to peer-based work. In this paper we are proposing the PrimeGame as a novel game-based programming tool which combines a wide variety of computer science skills in a small and manageable environment. The PrimeGame is a simple two-player board game which, however, is not fully-enumerable and thus can not be solved with pure computational power. Students are required to implement autonomous players for the game, in an attempt to model their own ideas on how to play the game best and to eventually beat the opponent. All submitted players will finally compete against each other in a tournament setting. As both, a simple player interface and a comfortable and fully functioning testing environment are provided, students don't have to care about the game engine, GUI components, or any other advanced aspects. They can instead focus on the design of suitable algorithms, as well as syntactically correct implementations thereof. The PrimeGame testing environment as well as the interactive graphical tournament framework including various on-line board presentations and ranking schemes are currently implemented as a Java project, thus student's players can easily be provided as Java classes. Common wrappers and interfaces for players developed in other programming languages are available as well. In developing PrimeGame players, students get early exposure to a combination of skills as necessary for all major computer science (CS) projects: from initial ideas and intuitive algorithms over studies of performance aspects, formal representations, implementation, up to evaluation and testing are all combined in a game project that can be individually completed within a few weeks time. Using the game in class, we witnessed an increased student interest in the subject matter, as well as a lot of fun among participating students. As the PrimeGame approach has so far been used and evaluated in undergraduate CS courses in not only European but also in African environments, it exposed different cultural attitudes towards competitive learning. While so far we have mainly focused on undergraduate aspects of the game, we are currently looking at advanced concepts which could be taught similarly well in graduate and post-graduate programmes. These include autonomous and co-operative agents, advanced data structures and algorithms, artificial intelligence and advanced numerical methods.