Virtual worlds modeling for Web deployment

We propose a design architecture for the graphics engine of a virtual reality system to be deployed over the Web. The graphics engine is decomposed into high and low level components. The high level component manages the hierarchical organization of the virtual objects, while the low level component acts as the interface to the graphics hardware and is mainly responsible for operations, such as vertex transforms and polygon rasterization. Several methods for geometry representation, compression and culling are compared and associated with different geometry types. The use of OpenGL as a general-purpose machine to implement mathematical operations is considered in the context of programmable shaders. Finally a prototype application for virtual worlds modeling is presented along with detailed discussion on technical and implementation issues.