1.2. Conceptual Model of VC Interpreter

The parsing, rendering, and interaction with a VC file is undertaken by the VC browser. Depending on the scene specification, scene rendering is oriented towards real-time interaction or off-line generation of animated scenes. Several viewpoints can be located into a scene, and the active viewpoint and hearpoint, for each user can be interactively selected. User interactions modify the view and audition of the active user with respect to its current viewpoint and hearpoint (the user associated with the selected window), and cause scene modifications according to the scripts embedded in scene nodes. Only one single viewpoint can be active for a user at one time whereas several hearpoints can be active simultaneously.

Figure 1-1 illustrates the conceptual model of VC interpreter. The scene description file is transformed into a Directed Acyclic Graph (DAG). It contains hierarchically organized nodes that participate in the multi-modal rendering of the scene together with internal and external message routing.