How to Define “Place” on the Web Paul P. Maglio Stephen Farrell Rob Barrett IBM Almaden Research Center 650 Harry Rd San Jose, CA 95120 USA {pmaglio, barrett, sfarrell}@almaden.ibm.com ABSTRACT We discuss the question of how to form communities to facilitate interpersonal interaction among World Wide Web users, and suggest five basic methods for associating web users based on their browsing context. Keywords Social navigation, place, world wide web. INTRODUCTION Three main issues facing designers of information systems incorporating social navigation are (a) which users should be made aware of which other users, (b) how should users be made aware of one another, and (c) how should these users interact. In this short position paper, we are concerned mainly with the first of these (but see [4,5] for some discussion of our approaches to the other two issues). In particular, we address the question of how to form communities on the World Wide Web (WWW) by considering a variety of ways to find boundaries of user groups based on user activity and web structure. More precisely, we consider WebPlaces to be locations in web space that are bound together by usage patterns, user interests, or web connectivity. Some places are defined by the structure of the web itself, and others are tied to the way in which users access the web. WEBPLACES In the trivial case, users might form a community when they are part of the same organization and thus use the same proxy server to access the web (see Figure 1). In this case, the users are likely to have similar interests or share similar views precisely because they are part of the same organization. This is like forming a community from people who routinely wait at the same bus stop but travel to different destinations. In a slightly more meaningful approach, WebPlaces might be defined by specific pages or sites (see Figure 2). Many such isolated places exist on the web (e.g., web chat rooms, game rooms, etc.), and one system, Virtual Places [6], was developed to add social awareness and interaction to individual sites or pages. The notion of place is limited by the authors of web documents to single documents or sites; place cannot transcend the accidental structure of the web. This is a little like forming a community from the people Figure 1: Users at same page, site, or domain. getting off the same bus at the same bus stop. A third method for associating users browsing the web is to find those who are exploring the same region, that is, who are visiting pages that are within a few links of each other (see Figure 3). For instance, one user who is looking for information on a particular disease might happen be visiting pages that lie within two or three links of pages others who are researching the same disease happen to be visiting. By taking advantage of the topology, well-defined communities or WebPlaces can be found [3]. This is like bringing together people who take different bus routes to the same destination. Communities can be tied to topics or categories of current Figure 2: Users exploring the same region interest. Traditional chat systems, such as IRC, as organized around content areas, such as Perl programming or soap operas. However, the web opens up new possibilities for connecting users who have similar interests, as determined by the topic areas they are currently browsing (see Figure 4). For instance, users searching for the same thing at different search engines might be considered to be in the same WebPlace. This might be like grouping together people who take different buses but wind up at similar kinds of destinations, such as museums or cafes. There are yet other ways to build user groups or communities. For instance, communities might grow from the repeated use of predefined paths that groups routinely Figure 5: Users with similar browsing histories. formal structures in which they move operate. We have briefly discussed how to form groups from users who are (a) visiting the same page at the same time, (b) connected to the web through the same firewall, (c) exploring pages that are closely linked together, (d) exploring pages with the same content, and (e) following similar trails through the information space of the web. We have begun to explore these notions of place in our prototype WebPlaces implementation [4,5]. ACKNOWLEDGMENTS We thank Tom Erickson for inspiration, and Andreas Dieberger for conversation. Figure 3: Users of the same proxy. follow using the WebPath Browser [2] or they might be cultivated from groups of sites where web users naturally congregate [1]. In the most general case, communities might be formed from web users with similar browsing histories (see Figure 5). This is like creating a community from people who routinely take the same buses. “Dynamic HTML” “Dynamic HTML” “Dynamic HTML” CONCLUSION The analogy between our notion of WebPlaces and a systems of buses and passengers is meant only to be suggestive. Our main point is that there are a family of methods for grouping users based on their activities and the Figure 4: Users exploring similar content. REFERENCES 1. Chalmers, M., Rodden, K. & Brodbeck, D. (1998). The order of things: Activity-centered information access. Computer Networks and ISDN Systems, 30. 2. Gibson, D., Kleinberg, J. & Raghavan P. (1998) Inferring Web communities from link topology. Proceedings of the Ninth ACM Conference on Hypertext and Hypermedia. 3. Gruen D. & Moody, P. (1998). Synchronous WebPath as a tool for leveraging expertise. Paper presented at Human Computer Interaction Consortium Conference, Snow Mountain Ranch, CO 4. Maglio, P. P. & Barrett, R. (1998). Adaptive communities and web places Second Workshop on Adaptive Hypertext and Hypermedia (Hypertext ’98), Pittburgh, PA. 5. Maglio, P. P. & Barrett, R. (1999). WebPlaces: Adding people to the web. Poster Proceedings of the Eighth International World Wide Web Conference, Toronto, Canada. 6. Shapiro, E. (1994). Virtual Places: A foundation for human interaction. Proceedings of the Second World Wide Web Conference.