Inhabited Digital Spaces
An Extended Abstract for a CHI 96 Formal Demonstration
Written in September 1995
343 Soquel Ave, Suite 70
Santa Cruz, California 95062-2035 USA
E-mail: via our Webmaster
Written September 1995 by Bruce Damer, Christina Kekenes, Terrel Hoffman, submitted to ACM CHI '96 and featured in Proceedings of the CHI '96 Conference Companion on Human factors: Common Ground, Demonstrations, ACM Press, New York, New York, ISBN:0-89791-832-0, pp. 9-10. Find the ACM Portal Digital Library Reference Here.
The emergence of standards such as Virtual Reality Modeling Language (VRML) has made shared, three dimensional virtual spaces available to the greater Internet community. When these spaces become inhabited by representations of people, often referred to as digital actors or avatars, a whole spectrum of social behavior will emerge. Prototypes of inhabited digital spaces have been hosted on the Internet since early 1995. Enough experience has been gained with these systems to produce an initial benchmark of their effectiveness from a user interface standpoint. Observation of social interaction in these spaces has also provided some interesting insights. A key finding is that there is a need for interdisciplinary collaboration between the technologists building digital worlds and specialists in community and social behavior.
virtual reality, social computing, electronic community
The inhabited digital space is a newly emerging medium of human interaction. This medium is born of the combination of two technologies, virtual reality and text-based chat environments such as Multi-User Domains (MUDs). Virtual reality has brought real-time three dimensional scenes to life on commonly available computer hardware. A growing number of MUDs and other text-based chat environments have brought a novel form of interactive community to users on networks such as the Internet. Several prototypes which combine elements of graphical virtual reality and text chat are now being hosted on the Internet. These prototypes allow users to manipulate a digital actor or avatar in the graphical space while communicating with others (also represented by avatars) using a text-based chat interface.
THE PROTOTYPES TESTED AND UNDER STUDY
Members of the Contact Consortium have been testing and studying several of these prototypes since early 1995. Prototypes tested include WorldsChat  (Figure 1) and Alpha World . Systems under study include Distributed Interactive Virtual Environment (DIVE) , Starbright World , and Worlds Away . Some of the prototypes examined are scheduled for release as commercial services. We expect that there will be an increased number of these systems available on the Internet in 1996. We hope that this benchmark will engage the interest of researchers and other contributors before the medium extends to the mainstream of computer users.
WorldsChat and Alpha World were tested extensively. To test, members of the Consortium would enter a pre-arranged area in the virtual world with agreed upon avatars and name badges. Cooperative walk-throughs of the spaces were tried to test the ease of navigation and following ability. Conversations were carried out between testers and other users in the environment. Interaction with novice and experienced users was carried out to gage their impressions of both the medium and its usefulness for interaction. Numerous screen shots were taken of typical social situations. Video was shot of one session in WorldsChat. Starbright World (still under development) was studied from videotaped sessions supplied by its designers. DIVE and Worlds Away were studied from the available literature and World Wide Web documentation and images.
Technical and user interface criteria
The prototypes had limitations in both performance and user interface. However, due to their use of a natural, point-of-view metaphor, they were surprisingly easy to navigate and engage other users in interaction. Very little in the way of user training was necessary. Often, veteran users would coach new arrivals. The technical and user interface criteria tested included:
- Appearance and rendering performance of graphics
- Avatar appearance and design (range of choices, badging, distance appearance, articulation or animation)
- Ease, options for avatar motion (degrees of freedom)
- Ease of communication (text chat, body language)
- Effectiveness of controls and services (doors, stairs, walkways, rooms, signage, coordinate readouts)
- Server crowd control effectiveness (speed of rendering other avatars, collision detection)
- Ability of users to alter the space (object moving and assembly, building construction, message posting)
- Special features offered in the world space (custom design of avatars, community newspaper, feedback facilities)
Observations of interactions
Observations of social interaction within these virtual worlds posed some challenges. We found that the lack of a number of important social cues lead to frequent misunderstandings between users. Observation focused on:
- Tracking threads of conversation between individuals
- Noting differences in interaction between the observer and newcomers or veteran participants
- Observing occurrences of aggression or cooperation, either communicated through text or by motion of the users' avatars
- Interaction with participants on their reactions to the environment and its effectiveness
- Observing the use of special terms in the environment
Figure 1. Scene from WorldsChat showing close range chat interaction with other users represented as avatars.
While the early prototypes suffered from both user interface and communication limitations, we were able to carry out meaningful interaction within these digital spaces. This medium may become effective as a tool for social interaction and cooperative work. Some initial areas in need of improvement are discussed below.
Improved graphics rendering and server crowd control speed would reduce the effort required to navigate and approach other avatars for interaction. Auditory cues when collisions occur would reduce confusion. Finally, a set of simple, standard interface cues for navigating in virtual space would be very useful. Some suggestion of how these cues might work is discussed by Slater, Usoh, and Steed in .
There are limitations that prevent the full expression of interaction, for example, because avatars are not unique to each user, it is difficult to distinguish between strangers and familiar users. Lastly lacking a means to identify who is involved in a conversation leads to frequent inadvertent interruptions. Some of these issues are addressed by Benford, Bowers, Lennart, and Greenhalgh in .
To facilitate improvements and overcome these limitations, we feel that there is a need for strong inter-disciplinary cooperation. Of particular importance is the involvement of social scientists and specialists in community in the process of design and testing of this emerging new medium of human interaction.
1. Smith, G., Living in the Virtual World: Phase One, in Popular Science (October 1995), Times Mirror Magazines: New York NY, p. 42.
2. Desmond, M., The Shape of Things to Come: Alpha World Brings a 3-D Virtual World to the Internet, in Multimedia World (September 1995), pp. 16-17.
3. Benford, S., Bowers, J., Lennart, F. E., Greenhalgh, C. Managing Mutual Awareness In Collaborative Virtual Environments, in Proc. VRST'94 (Singapore, 1994), World Scientific: Singapore, pp. 223-235.
4. Spielberg Launches Virtual Environment, in Southern California Micro Publishing News (July 1995), Micro Publishing Press: Torrance CA, pp. 1, 37.
5. Worlds Away, in VR World (July/August 1995), Mecklermedia Corp: Westport CT, p. 9.
6. Slater, M., Usoh, M., Steed, A., Steps and Ladders in Virtual Reality, in Proc. VRST'94 (Singapore, 1994), World Scientific, pp. 45-54.
(c) Copyright on this material is held by the authors (1996).
Back to Papers Index
Back to Consortium Home Page
© 1998 Contact Consortium, All Rights Reserved.