My virtual world is a complete room with 4 walls and a floor. There is a divider in the room so that the visitor is forced walk around the divider to see the diamond cube floating in midair. 

• How to configuration files work in simulation and real (with actual tracker and HMD) mode;
• How to setup the VR-Juggler development environment in Visual Studio (finding libraries and include files takes time);
• How the application connects to the rest of VR-Juggler; and
• Understand the tracking model of the head, wand, and camera (simulation mode).

Once I understood VR-Juggler and the “torus” example I could develop my worlds of cubes. The world is developed with OpenGL. My building block is a 1x1unit (I think unit=meters) cube. A wall (drawwall()) consists of the cubes in a checker board pattern. The walls can be made variable sizes, duplicated, rotated, and translated in the virtual world. In the process I learned the following:
 

• What the orientation and scale of the virtual world should be to fit into the real-world lab. The origin of the world is at the EVE cabinet. Dimension XY are along the floor, whereby Y is from the origin to the Latency1-cs computer. The Z dimension is up.
• Standing “in” the virtual-wall and looking along the long surface can verify the placement. 
• Relearned to program OpenGL. This is a useful tutorial site: NeHe's OpenGL Tutorials,  
• I learned that the checkerboard wall pattern is more appropriate than a solid wall. It is difficult to notice motion when looking a solid colored wall. It was obvious to see motion when looking at the checkerboard-patterned wall. 
• The frame rate of the virtual world seems slow as the viewed images “jump” as the viewer’s head turns.

The code is here

• torusApp.c -- The code where the virtual world is drawn (see drawroom()).
• The rest of the code is like the torus example. 

• Texture mapping. I wanted to use texture mapping to make objects standout. I abandoned the idea after I could not find a standard way to load the texture’s bitmap. Choosing a non-stand method is OK, but I would have to get it to work on two computer systems.
• I am experimenting with spatial audio using OpenAL (Creative labs). In a non-standard way I was able to integrate an OpenAL into the VR-Juggler example. Sound can be generated, however, is incorrect as I have not yet aligned the audio virtual world to the lab coordinates.