Widening the Net: Workshop Report on the Theory and Practice of Physical and Network Communities
Steve Whittaker, Ellen Isaacs & Vicki O'Day
Vol.29 No.3, July 1997
Activity and object building
Orientation and help
Maintenance and growth
About the Authors
We report the main discussions and highlights of a workshop held at the Computer Supported Cooperative Work conference in Boston, on November 16th and 17th, 1996. Here we present the organisers' perspective of the workshop: the position papers along with one participant's (Gail Rein) perspective are also reported in the ACM SIGGROUP Bulletin (April and August editions). The word "community" is now a popular and loaded one in the context of the Internet. A number of independent factors seem to be contributing to the popularity of the term. On the one hand, certain classes of multi-user application are growing in popularity (e.g. MOOs, MUDS, virtual worlds). These are specifically intended to provide environments in which large numbers of users can "live", work and play together. The designers of these systems often label these as "community systems". There are also recent attempts to provide existing geographically-centred communities with electronic support and services (Agre & Shuler, 1996, Carroll & Rosson, 1996, Mukhopadhyay et al., 1996). In addition, social scientists who have studied communities in the physical world are beginning to look at established on-line systems such as UseNet, MUDs, Bulletin Boards, chat rooms and Internet Relay Chat to determine whether the behaviour of users of these systems does indeed concur with their theoretically derived definitions of community (Baym, 1995, Kollock & Smith, 1996, Mynatt, et al., in press, Reid, 1991, Wellman & Gulia, in press). Finally, there are futurists and social critics predicting widely different implications arising from the growth of these virtual "communities". These new communities will either undermine or revitalise their real-world parallels depending on the writer's bias and sensitivities (Mitchell, 1995, Rheingold, 1993, Stoll, 1995).
There are good reasons why the issue of community and systems to support it should be of interest to CSCW researchers. In the past, CSCW systems have tended to focus on the support of relatively small groups in work settings carrying out work-related tasks (Greif, 1988). In contrast, the goal of these new "community" systems is to support groups of hundreds or even thousands of interacting users, who can contribute to the construction and maintenance of their own virtual environments and sometimes set collective policies for their governance. Some of these systems also allow embodiment and provide extensible spatial environments for navigation and object construction. Given the interest in this new class of system and these differences from current CSCW systems, it is important that the field arrive at some insights into the theory and design principles associated with this novel class of system.
The workshop aims were to (a) identify outstanding theoretical issues in our understanding of physical and network communities, and (b) develop a set of design requirements and principles for building community systems. The goal was to bring together participants from different disciplines and research backgrounds to promote cross fertilisation between social theorists and community system builders.
Our first sessions addressed the issue of definition, by a process of generating positive and negative exemplars of community. There was a remarkable diversity of opinion about the positive examples, with suggestions ranging from swarms of bees to "doggie parties" -- pet owners who get together to talk about and play with their pooches. More classical (non-animal) examples of community included an English town, or a parent participation group. We were able to identify a set of key dimensions of community, by analysing and contrasting the underlying characteristics of both positive and negative examples. We settled on an approach of defining the concept by "prototypical attributes", so that communities with more such attributes were clearer examples of communities than those that had fewer.
Core attributes were:
- members have some shared goal, interest, need, or activity that provides the primary reason for belonging to the community
- members engage in repeated active participation and there are
often intense interactions, strong emotional ties and shared activities
occurring between participants
- members have access to shared resources and there are policies for determining access to those resources
- reciprocity of information, support and services between members
- shared context (social conventions, language, protocols).
Less central attributes were:
- differentiated roles and reputations
- awareness of membership boundaries and group identity
- initiation criteria
- history and long duration
- events or rituals
- shared physical environment
- voluntary membership.
There then followed a series of demonstrations of the Pueblo MOO (O'Day et al., 1996), and several graphical virtual worlds, including WorldsAway, AlphaWorld, WorldsChat, Starbright and Onlive, (for more information see Contact Consortium Web page, http://www.ccon.org). All these have been described as "community systems". We identified a number of key differences between these worlds. One difference is the extent to which users are visually embodied. For example, in MUDS users communicate using text only, and sense of location and navigation is achieved through textual description and commands. In contrast, in Graphical Worlds, users have visual manifestations as avatars and environments are depicted rather than realised in text. A second major distinction between the worlds lies in the importance of object building and object construction. In MUDS and certain of the Graphical Worlds such as AlphaWorld, there is support for object and environment construction. In contrast, the focus in WorldsAway (and other worlds we did not examine such as The Palace) is more on interpersonal communication and interaction than in complex environment construction. In addition to these differences in primary activity, we also identified differences in the character and "feel" of the different systems depending on the nature of the participants and their interaction topics and conventions.
We also discussed the commonalities between phenomena observed in virtual worlds and those occurring in physical communities. People pointed out the capacity of these systems for engendering strong emotional behaviours. Virtual worlds have seen on-line romances, sometimes leading to real-world meetings and long-term relationships. On the negative side, virtual worlds have also suffered from some of the same problems as real-world communities, there have been power disputes, as well as personal and property violations.
We then examined the central characteristics of these virtual communities to determine the overlap with our earlier discussion of community definitions. We hotly debated which of the core aspects of our definitions were critical in design. Some participants felt that the central issues surrounded support for social interaction: namely how we promote intense interactions with others. Others felt that the crucial issues concern collective action and the ethos of the system: given that users share resources, what policies do they have to regulate access to those resources, what conventions are in place for interaction with others and how do users learn these policies and conventions?
Without reaching definitive agreement on this issue, we identified a number of key areas that need to be addressed. Some of these apply to fundamental behaviours in these systems, such as interaction, self-presentation, actions and movement:
a. Conversation -- given the centrality to community of repeated interactions between participants, how are these supported, especially given the fact that there are often large numbers of participants sometimes holding multiple simultaneous conversations?
b. Identity and self-presentation -- these are inherently social environments, so how in a network community do users provide others with social cues about their personality and characteristics? This is a critical issue because interactions often occur between people who are strangers. In addition, in some worlds, role-playing and exploration of identity is the primary focus of many participants.
c. Activity and object building -- many of these environments allow users to construct or modify objects and make their own personal spaces, but what are the rules that govern this and what are the tools and materials that support it?
d. Navigation -- for worlds with a spatial component, how do users move, determine where they are and where they can go to next?
We debated how the realisation of these behaviours and the consequent design choices differ according to the nature of the underlying technical system. In Internet Relay Chat for example, there is no embodiment, self-presentation is limited to a choice of name ("handle"), navigation consists of specifying one of multiple channels, and there is little capacity for constructing objects. Given the lack of embodiment in IRC, there was some discussion about the appropriateness of the term "community" to IRC behaviours, but this was countered by the observation that repeated interactions, strong ties, and social conventions are pervasive on IRC.
The other central design issues concerned the longer term ("lifetime") aspects of the system:
e. Culture and policies -- how are shared resources managed, how are individuals and individual property protected, how is reciprocity promoted, how are newcomers encouraged and assisted and how do they know the norms of the world when they first enter it?
f. Change and growth -- these worlds are organic, so given that neither the structure of the world nor its set of inhabitants is preordained, how is change managed?
Our final session involved taking these general design requirements and trying to provide a set of specific guidelines to address them. We examined different system exemplars (namely MUDs, Internet Relay Chat, graphical virtual worlds and UseNet newsgroups). For each we considered different contexts, namely work, social and interest-based settings, to consider design choices and evaluate the trade-offs between them. We came up with the following set of guidelines:
Users should be able to...
- Tell who is speaking
- Provide feedback to a speaker verbally or non verbally
- Follow a conversational thread when there are multiple speakers
- Backtrack on a conversation
- Know if conversation is being logged
- Signal which conversational thread they are attending to
- "Hear" everything that is said (locally) with enough time to process it
- Communicate with different "voices" (e.g. say, emote, think, change real voice)
- Indicate level of engagement (e.g. be able to indicate when otherwise engaged in real world conversation).
Users should be able to...
- Have control over their own avatar, identity, or presentation
- Know how they appear
- Express subgroup identity (e.g. via status symbols)
- Know when others are playing a role vs. being themselves.
Users should be able to ...
- Control their own objects and conduct actions on them
- Be aware of in-world changes but not out-of-world changes (e.g. a change in server should not be signalled to participants)
- See the same changes that others see
- Be aware of possible agents of change and be able to find out who brought about changes
- Determine rules of object ownership regarding: deletion,
addition, changing, copying (trademark issues), giving access, sharing
- Modify existing objects with owner's permission
- Create new objects from existing ones
- Contribute objects to the public domain to be used or adapted by others.
Users should be able to...
- Tell where they are
- Get back to "ground zero" easily
- Determine consistent navigation rules (i.e. the rules of navigation should be location independent)
- Easily gain a sense of the world layout (e.g. from maps).
Users should be able to ...
- Do meaningful activities at the outset without help
- Be rewarded for helping others learn the rules and get oriented
- Gain access to formal helpers and help documents
- Determine the norms for helping
- Easily install, and maintain the software.
It is useful to ...
- Have explicit rules for membership
- Provide support and rewards for leaders, good examples and altruistic behaviour
- Provide mechanisms for "repossessing" content
- Allow, and build in, "interesting" changes in structure (e.g. add "weather" or natural disasters)
- Allow safe ways to try out changes, and allow these to be reversed easily
- Minimise surprises in work setting
- Support development of individual credibility, in part by enabling persistent identity
- Provide artifacts to show, record history
- Allow people to bring in stuff from the real world
- Provide assurances for security, reliability
- Provide ways to resolve conflict and sanction antisocial behaviour
- Enable users to modify the structure of the world
- Maintain peripheral areas (web sites, maps, schedules) to attract newbies.
In all areas, virtual communities and worlds should enable things that are not possible in the real world, e.g. thinking versus saying, being in two places at once, telepathy, teleporting, sharing a personality.
It was clear from our discussions that the principles for conversation, identity, navigation and object building are more specific than those for orientation, help, maintenance and growth. We concluded that part of the reason was that the first four topics address concrete issues in the user interface and basic user actions, where the design space and options are reasonably well-defined. In contrast, orientation, help, maintenance and growth concern the ethos and policies that are associated with the community. These depend partially on judgments made by the community designers at the outset, but more importantly on decisions made by the participants of the community as it evolves. It is therefore harder to legislate for these, given that much will depend on the nature of the community, its context of operation and evolving ethos.
Mark Ackerman, (UC Irvine), Annette Adler (Xerox Systems Architecture), Daniel Bobrow, Xerox PARC), Joern Bollmeyer (Paderborn University, Germany), Bruce Damer (Contact Consortium), Paul Dourish (Apple Research Lab), Thomas Erickson (Apple Research Lab), Mark Jones (Andersen Consulting), Jim Larson (Intel Architecture Labs), Jin Li (IBM Software Solutions Toronto Lab), Wayne Lutters (UC Irvine), Ioannis Paniaras (University of Art and Design Helsinki, Finland), Gail Rein (Xerox Systems Architecture), Duncan Sanderson (University of Brighton), Jeff Sokolov (GTE Labs), Konrad Tollmar (KTH/IPLab, Stockholm, Sweden), Catherine Wolf (IBM Watson).
Agre, P. & Shuler, D. (1996). Reinventing technology, rediscovering community: critical explorations of computing on social practice. Ablex, Norwood, N.J.
Baym, N. (1995). From practice to culture on Usenet. In S. Star (Ed.), The cultures of computing. Blackwell, Oxford, UK.
Carroll, J. & Rosson, M-B. (1996). Developing the Blacksburg Electronic Village, Communications of the ACM, 39 (12), 69-74.
Contact Consortium, http://www.ccon.org.
Greif, I. (1988). Computer Supported Cooperative Work. Morgan-Kaufmann, San Mateo, CA.
Kollock, P. & Smith, M. (1996) Managing the virtual commons. In S. Herring (Ed.), computer mediated communication: linguistic, social and cross-cultural perspectives. John Benjamin, Philadelphia.
Kraut, R., Scherlis, W., Mukhopadhyay, T., Manning, J. & Kiesler, S. (1996). The Homenet field trial of residential Internet services. Communications of the ACM, 39, (12), 55-63.
Mitchell, W. (1995). City of Bits. MIT Press, Cambridge, MA.
Mynatt, B., Adler, A., Ito, M., & O'Day, V. (in press). Designing virtual communities. To appear in Proceedings of the Conference on Computer Human Interaction, 1997.
O'Day, V., Bobrow, D., & Shirley, M. (1996). The Social-Technical Design Circle, Proceedings of the Conference on Computer Supported Cooperative Work, 160-169.
Reid, E. (1991). Electropolis: communications and community on internet relay chat. Unpublished masters thesis, University of Melbourne.
Rheingold, H. (1993). The virtual community: homesteading on the electronic frontier. Addison-Wesley, Reading, MA.
Stoll, C. (1995). Silicon snake oil. Doubleday, New York.
Wellman, B. & Gulia M. (in press). Netsurfers don't ride alone: In S. Kiesler (Ed.), The culture of the Internet.
Steve Whittaker is a researcher at ATT Labs, working on computer-mediated communication. He is interested in the design and impact of novel communication technologies on individuals and distributed organisations. Steve also researches user interfaces for browsing speech data.
Ellen Isaacs is a user interface designer at Electric Communities, a start-up company building a platform for, and instances of, on-line virtual worlds. Her previous work focused on the design and use of multimedia-based communication and collaboration tools to support distributed groups, with a focus on awareness and enterprise-wide communication.
Vicki O'Day is a researcher at Xerox PARC. She participates in, studies, and contributes to the design of an educational MOO called Pueblo. She is especially interested in the ways technical mechanisms and social practices develop and evolve together in successful network communities. Her past work includes ethnographic studies of librarians and other information experts and the design and implementation of new technologies for sharing information.
Steve Whittaker, ATT Labs-Research, 600 Mountain Ave., Murray Hill, NJ 07960, USA.
Phone: +1 (908) 582 6551
Ellen Isaacs, Electronic Communities, 10101 De Anza Blvd., Cupertino, CA., 95014, USA.
Phone: +1 (408) 342 9540
Vicki O'Day, Xerox PARC, 3333 Coyote Hill Rd., Palo Alto, CA., 94304, USA.
Phone: +1 (415) 812 5051
Vol.29 No.3, July 1997