I taught Richard through his degree and supervised him through his PhD along with my colleagues David McSherry and Michael McNeill. This week he got to the end of his long journey at Coleraine by successfully defending his PhD thesis within the area of interactive adaptive digital storytelling.
During his research he developed a storytelling system called MIST. For me this research project was very interesting and Richard made several very useful contributions to adaptive story planning – particularly in relation to plan failure detection and re-planning within the context of persistent, multi-player online worlds. My motivation for being interested in this work is partly that I’m dissatisfied with MMO game design over the past decade – and I’ve played almost all the big games. What we set out to achieve on Richard’s project was to develop a way for story-lines to be more emergent, for player and NPC behaviour, actions and choices to have an impact on the story that other players experience. Existing games use variations of instancing and phasing along with fixed quests to provide stories, and to a large extent players all experience the same story. Personally, I have become bored with this.
What we achieved on Richard’s project was to remodel HTN planners within the MIST architecture so that the system can provide abstract story plans, which have many instance permutations within a dynamic, non-deterministic game world. A story manager, much like a dungeon master in D&D games, then ensures that stories are valid and structured properly.
Richard’s thesis needs a few minor corrections before it is ready but we have 2 published papers that provide some the detail of the MIST story system: