Usability of web-based VR - introduction

Here, we want to share some ideas and experiences on three-dimensional, interactive applications on the World Wide Web.

A few years ago, several people were rather skeptical about web-based Virtual Reality, see Jakob Nielsen (1998), 2D is Better Than 3D, for an example. This skepticism was rooted in the discrepancy of what hardware is available in specialized laboratories, namely sophisticated and thus expensive projection systems and interactive devices, compared to what WWW users have at their disposal. This raised the question whether under the latter conditions at least some degree of immersion can be realized, as the pivotal quality criterion of VR applications.

Meanwhile, creative developers have shown that users can get absorbed even by rather simple computer games, providing evidence that such a technology driven approach to immersion is too narrow. Moreover, a wide spectrum of web-based VR-applications has evolved, including data visualization, simulation and modeling, virtual prototyping, simulator training, education, communication, marketing, healthcare and telepresence. And finally, technical bottlenecks like low bandwidth and graphics board performance are getting less and less relevant.

Now, the question is not whether or not but rather how to design VR applications for the WWW. As this field is quite new, compared to conventional HTML applications, developers are very creative in designing the interface. On the other hand, this diversity results in several and quite severe usability problems. E.g., each time you start a VR application you need to read and memorize instructions on how to handle it, and frequently you give up already at this stage because of the required time and effort.

Usability is a quite enigmatic and complex concept, but it seems to be rooted at least in the following two general principles:

• Affordances: James J. Gibson (1977, 1979) coined the term to refer to the actionable properties between the world and an actor. Or - in more plain words - the environment provides cues of what are useful actions in a given situation, and the human perceives and reacts to these cues in a rather automatic way, based on innate knowledge. An example is the attractiveness of buttons and switches to babies; one of their first and most favorite toys is the remote control of the TV.
• Consensus: However, Gibson's concept is only one - and constructivists would say the smaller - part of the story. For example, it is neither natural nor self-explanatory that a hyperlink is represented as colored and underlined text. Rather, its functionality is based on an agreement between the sender (developer) and the receiver (user). And, unfortunately, there are cases in which such agreements are ignored - have you ever been clicking on underlined text, finding out that it is just used for accentuation purposes? - making applications unusable.

We speculate that the success of conventional HTML applications is mainly based on establishing social agreements on how to design and use them. This process is backed by numerous guidelines and checklists, one example are the Research-Based Web Design & Usability Guidelines.

Unfortunately, none of them covers VR applications. In addition, the Web3D Consortium and developer centers for proprietary viewers (see, e.g., Adobe) do not provide support on usability questions, but rather focus on the technical realization of certain VR features und applications. As a result, this field is rather ruled by anarchy - consensus on the essential principles of user interface design still needs to be established, to open up the full potential of VR in the context of web-based applications.

On the following pages we will discuss some general aspects of usability, namely the selection of an appropriate 3D format, the realization of navigation options, and the proper representation of interaction functionalities. These discussions should not be viewed as providing the 'ultima ratio' - rather our intention is to contribute to the development of some basic standards, on the background of the experiences that we acquired with our own VRML and Shockwave applications in the field of eLearning.