Towards better personas in gaming : Contract based expert systems

Recent evaluations of Procedural Content Generation (PCG) methods have examined the use of personas as part of there evaluative functions. Personas, models of a user, have a number of complaints leveled against them by researchers as to their use as design tools: that there is poor empirical proof that they improve the design process, poor definitions lead to them not modeling actual uses, and they do not give any quantitative information for evaluation using a scientific approach. Examined is a framework in which a persona is not just defined as a model of a user, but is a contract list of player goals and actions available in the game. By defining a persona as a goal driven AI, it joins the idea of personas with a more qualitative evaluative method as well as not allowing developers to apply their own biases. Further, it allows for a persona to work beyond the design stage of the development, and into the maintenance stages. Taking into account game-play telemetry, it allows for personas to be used as tools for lessons learned for future developments.