Integrating semantics and procedural generation: key enabling factors

Manual content creation for virtual worlds can no longer satisfy the increasing demand arising from areas as entertainment and serious games, simulations, movies, etc. Furthermore, currently deployed modeling tools basically do not scale up: while they become more and more specialized and complex, they strikingly fail to assist designers in concentrating on their creative tasks. This paper presents the main concepts of what we designate declarative modeling of virtual worlds, a novel modeling approach that allows designers of virtual worlds to concentrate on stating what they want to create instead of on describing how they should model it. We discuss the two main characteristics of this approach – semantics and procedural generation – and describe how they mutually reinforce, support and amplify designers' efforts, thus empowering their creative skills. We conclude that this declarative approach provides designers with the productivity gain of procedural generation techniques, while still allowing for abundant control and flexibility. In addition, it significantly reduces the complexity of virtual world modeling, making it accessible to whole new groups of users and applications.