Introduction

It is fundamental you thoroughly develop a concept before you begin working. A clear map of what you want to show and how you want your users to interact with your data are fundamental to any electronic visualization project. Your concept will evolve and possibly change as your project begins to take shape, but it is a good idea to start with a basic problem such as a transit oriented redevelopment, decide how you want to represent the physical environment, such as in a density study, and how the environment will change over time (models increase in bulk from original to proposed as the user enters within 10 feet of the building). Layering over of the basic problem, you should also consider how enhanced visual presence or immersive interaction will aid the project. I encourage you to consider pushing your concepts beyond aesthetic observations by utilizing interactivity.

Choosing between VRML and the CAVE is one of the first decisions to make. You must consider your audience, accessibility issues, and the complexity of your project. Fully define the concept before working to assess your project timeline. Generally, I would double the amount of time you think it will take you, but for CAVE programming, I might even be inclined to triple it. VRML projects for the web will not take you as long because these projects require significantly less steps. Novices to 3D interaction and world creation will have innumerable stumbling blocks, however you gain a plethora of experiences and exposure to key ingredients that create a successful interactive piece.

CAVE Projects

It is important to be able to justify a CAVE project. If you could easily accomplish the same results on the Desktop, you might want to reconsider a CAVE project. The dilemma with the CAVE is the creation of intuitive navigation, interaction, and accessibility. Usually, a designated "driver" shows a CAVE piece to an audience, while a VRML piece can be shown over the web with instructions for the user. The advantage to using the CAVE is the obvious immersion of the user but also the CAVE can be networked with other CAVEs. Users from thousands of miles away can enter into the same virtual environment simultaneously and can talk to each other while interacting with the same objects. Rem Koolhaas used the CAVE to have a design charette for the New Campus Center at the Illinois Institute of Technology in Chicago without leaving Holland.

Most user interfaces in the CAVE, the way in which you manipulate the environment, are programmed. The CAVE only comes with the default walk mode. You must program and plan for all other types of navigation (such as flying and inspecting) as well as the interactivity. Your project will introduce you to a world of computer interfacing beyond basic navigation issues. It might be helpful to research how navigation and interactivity has been addressed in the past to help you make informed choices.

For the Hazel Crest Planning project, an interface for moving buildings was constructed that enabled the user to click a button on the wand and the world would change. Like a garage door, the world slips from underneath and around the user to be displayed on one wall of the CAVE in aerial view. The user can point at buildings and rearrange them in this perspective. This was designed around the nature of the transit oriented project which included replacing most of the buildings within a Metra corridor. While standing in front of a particular building, the user was presented with a series of three dimensional buttons that enabled him to scale, rotate, and swap buildings. The user is presented with the existing conditions and can replace the existing condition with the proposed development, as well as scale and rotate the new building.

This experimental Hazel Crest piece was both a success and failure. The idea of restricting views to two dimensions when the user wants to move buildings worked wonderfully. Within a three dimensional environment, it is very difficult to stand in front of a building and move it to a more desired location. The buttons to swap, rotate, and scale could have been better integrated into the built environment without taking up physical space in the simulation. Way finding was not a concern because the corridor only consisted of several blocks. Planners did not have time to come to the CAVE. Thy only interacted with the VRML piece minimally. Due to timeline constraints, the planners developed plans about the corridor without the VRML simulation having any affect on their decisions. The simulation was under construction right up to the community meeting; its only use was to convey the plans to the community. The final community meeting was successful and the audience asked the planners to "drive" the corridor to address questions immediately. It was a successful visualization tool but not a tool the planners actually used to make informed decisions.

Image from Hazel Crest TOD Project, done in both VRML and the CAVE, Crystal Wilson, 2002. This image shows the three dimensional button to enable the user to switch between new and existing buildings.

VRML Projects

VRML projects present similar navigation issues, however, the plug-in viewers for the browser have inherently more functionally built in than the CAVE wand. It is easier to navigate a VRML world. Added functionality can be added to both the CAVE and VRML. In the Kenosha VRML project, a two dimensional map can be called up from the bottom of the screen. The user's position is tracked on the map for way finding purposes. This method is one way to expand the navigation of the virtual environment.

Below is a CAVE Versus VRML overview. You can make an informed decision about which path to take when you consider what you want to do in your virtual world, the project timeline, and audience.

* Inherently Networked means that more than one person in different locations can enter into the same environment together.