Declarative Scenes Modeling With Dynamic Links And Decision Rules Distributed Among The Objects

Declarative modeling offers to the user the possibility to describe a scene without giving numerical data, but instead to express the mental image of the scene in almost natural language. The translation of the natural language like commands into accurate numerical data is delegated to the system.In this paper we present a declarative modeler for three-dimensional scenes for use in image synthesis. The design of the modeler is object oriented. A graphical interface allows the user to specify high level commands. The events manager and the interactions model define the user-system dialogue. Dynamic links and decision rules are distributed among the objects. The objects can decide for themselves the appropriate actions necessary to keep the scene coherent.