DIVER: An Immersive, 3D Graphics System

DIVER (Distributed Virtual Environment Rendering system was developed in the early 1990's at the University of Virginia. At that time two separate $100,000.00 SGI rendering engines drew graphics -- one per eye while a Polhemus tracker used magnetic fields to determine where the user was located. The user wore a 10 lb (bowling ball) headset that used special optics to project the images into each eye. It was DIVERs job to provide the graphic API and to do the distributed rendering.

Unlike the static OpenInventor file format, DIVER used a dynamic rendering tree. The programmer could create one or more camera objects and place them anywhere in the tree. The camera would inherit the position of the parent object in the rendering tree. This made it very easy to create a "Bird's Eye View" simply by attaching the camera under the bird in the scene. You could also have the camera pointAt and moveWith objects in the scene, allowing the camera to track.

This idea was extended to all objects in the scene. In otherwords, moving an object from one parent to the other in the scene graph did not change the object's position in space, it just changed the parenting. Of course, once the parent object moved, the child object would move along with it. A ClearMatrix call would allow the programmer the means to also move the object underneath the parent object in 3D space as well.

DIVER was the pre-cursor to Alice now at CMU.

Papers


©2005 Rich Gossweiler, all rights reserved.