Published: March 15, 2001.
Co-operation Metaphors for Virtual Museums
Thimoty Barbieri, Paolo Paolini, Dept. of Electronics and Computer Science, Politecnico di Milano, Italy
Cooperation Metaphors are sets of rules to support interaction and collaboration between users who want to explore complex content and information together. The rules determine how the collaborative community can be created and managed, how members of the community can operate on their own or can cooperate with other members. Different types of situations, tasks and user roles determine different behaviours and therefore need different metaphors. The paper will present the general issue, describing in general what cooperation metaphors are and how they can be defined. A number of examples, related to typical activities for virtual museums (i.e. museums on the Web) will be used in order to give a practical understanding of what cooperation is, or can be. The paper will also argue that virtual metaphors lack some of the features of real-life cooperation, but, on the other hand, can also offer unexpected, powerful and effective possibilities not available in (traditional) real-life. Finally, the paper will describe how collaborative activities for virtual museums can be implemented using today's tools and applications for collaboration on the Web; such as Net2gether, Microsoft Research's Virtual Worlds, and the WebTalk series.
Keywords; Collaboration, Virtual Museums, CSCW, CSCL, Virtual Reality
Introduction and Motivations
In most cases, users today access the shared space represented by a web site in isolation, with little awareness, or no awareness at all, of what other users are doing "at the same time". In a truly cooperative environment, by contrast, different users try to accomplish something together, accessing the shared space simultaneously and trying to collaborate to reach their goal. Several different goals may be the subject of cooperation: teaching/learning, shopping, design, supervising and monitoring, consulting, etc. Several different behaviours are associated with the notion of cooperation, but we should distinguish between two extremes:
One basic observation informing our work is that in the real world most activities are performed, if possible, in cooperation. Cooperating while trying to accomplish something very often appears more interesting, more engaging, more amusing or simply more efficient,
This initial observation is coupled with the fact that virtual shared environments have existed for a number of years. Initially they were text-based (IRC-II, MUDs, MOOs), allowing users to exchange typed messages, but they have evolved (Blaxxun Community, Virtual Worlds) introducing 3D graphics, audio, and other advanced features. In these "virtual worlds", as they are usually called, users are represented by virtual objects, called "avatars"; users can "move around", cooperatively interact with the virtual world and its objects, "talk" each other, etc.
These worlds allow a great deal of interaction among users, but they lack some of the features that we feel are crucial for an effective collaborative visit to a Virtual Museum:
User interaction seems to be the goal, rather than the means to accomplish something. Users amuse themselves by interacting, but they are not trying to get any specific result out of their experience. Visitors to a Virtual Museum, on the other hand, share the goal of better understanding the content and the background of the museum.
The amount of information being exchanged among users, or being made available by the world, is relatively small and loosely organized. Sophisticated museum sites by contrast convey a large body of sometimes very difficult knowledge.
The cooperation is loose, in the sense that users are free to behave as they want, with few constraints or enforced patterns. Our experience, however, shows that cooperation must be organized and follow precise guidelines in order to be effective.
Based on this we have been working on the following ideas since 1998:
With the above prerequisites, a development environment, WebTalk-I (WebTalk, Barbieri, 1999) was developed and a few applications were developed. The best known of these applications was developed for the Museum of Science and Technology of Milan. The application, named "Virtual Leonardo" (Paolini, 1999), allowed users to co-operate while accessing a site with a number of pages describing different machines invented (designed but never built) by Leonardo da Vinci. Virtual Leonardo was presented in previous editions of Museums and the Web.
Virtual Leonardo allows users to visit the museum together, exchange opinions, and interact with the world and with each other in a number of ways: interactive gateways allow the users to visit the pages of the Website though standard Web technology.
Our experience with Virtual Leonardo, reinforced after collecting usage data and impressions from the users (Barbieri 2000), convinced us of the necessity to support the creation of such collaborative environments with a theory of design patterns. These are the subject of this paper.
Designing for Collaboration
Within a cooperative environment the goal, for a user, is to accomplish something specific (e.g. to get work done, to learn something, to buy something, etc.) through cooperating with other users. The other users can be at the same level of knowledge and power, or can be more expert and/or have greater control over the application. A number of visitors in a museum, for example, can be at a comparable level, while a museum guide is a user with more expertise and with the power to "take" the group of visitors around.
When the users interact with each other, they follow patterns of interaction that we call "cooperation metaphors". The goal of this section is to analyze the general features that cooperation metaphors must satisfy, in order to be effective, and at the same time engaging and compelling. The discussion that follows is based upon literature and our experience gained with WebTalk-I (Barbieri, 1999) and the deployment of applications.
First of all, in a cooperation environment, it is important that all actors share the notion of a common state, bound within time and space. The intensity with which this feeling is conveyed to all users determines the level of awareness (Gutwin, 1997) within the environment. The higher the awareness of the shared state, the better the cooperation between users: thus awareness is the primary effect we must create to allow users to work and discuss together. In a virtual space described in three dimensions, users can experience perceptual stimuli that are more similar to everyday life. By processing and understanding these stimuli, they can interact with the environment and the other actors, increasing their awareness, and creating in turn new stimuli for the others. In a collaborative system, this leads to the phenomenon of prediction; that is, participants can anticipate, within a short time, the actions of the others, because the number of informational details they have regarding the space and time they share with others (in one word, their awareness of the system) can lead them to such conclusions. (In this respect, for example, the "slow motion" of avatars in a 3D space can be more effective than fast motion or jumps, since the other users can better predict where an avatar is going if it is moving around slowly). Other important consequences of awareness are the possibility of learning indirectly from other people's activity, or of gathering information from the state or behaviour of the artefact in the shared environments, or the possibility of using gesturing or indirect communication forms to support one's communication (e.g. "Follow me this way", "Take this..", etc.) when coordinating multiple actions for which more users are needed.
The problem in conveying the correct form of awareness is often application-related, meaning that not all kinds of collaborative information are always needed to perform a task within a determined environment. That is why before deploying a collaborative system it is important to design in advance the means by which information will be presented in three-dimensions (Spatial Patterns), and to determine precisely the rules by which users are able to cooperate with each other, and with the environment, and to enhance their awareness during their activity. We call this set of rules cooperation metaphors.
In our vision, a cooperation metaphor (Barbieri, 2000) is a set of basic rules that describe the different modalities of interaction between users and between users and their environment. These rules encompass various aspects, such as the way users can gather in groups to talk to each other or navigate virtual space, or how visualization of the state of the artefacts and of the avatars (the figurines which represent the current position of each user in space) is to be performed. By making decisions about all these aspects, it is possible to define a global description (a metaphor) of the possible forms of collaboration of users within a system. In the following section, we present an abridged list of these small rules, which can be freely selected to form any metaphor that is felt suitable to regulate collaboration within a certain application.
In addition, the design of the application also calls for deciding which spatial patterns are to be followed. Such decisions include whether the world will be represented in an abstract way or by mimicking reality in a simplified way (Bridges and Charitos, 1997, Benford and Snowdon, Hearst, and Karadi 97). Moreover, we must decide how to represent, in a consistent fashion, important visual elements that aid the navigation in space of the users, such as places (Cerulli, 1999), thresholds between one space and the other, landmarks to assist users in finding their own routes within the space (Charitos and Rutherford, 1997), spatial hyperlinks to jump from one point of the space to another (Campbell, 1996, Charitos, 96), and so forth. We will not investigate this complex issue in detail in the present paper.
Our (limited) experience in designing cooperative applications has shown that although the basic cooperation metaphors are the same, the way to actually implement them or to assemble them, changes from application to application. The role of the designer of a collaborative environment, therefore, is not solely limited to gathering contents and drawing 3D geometries, but includes conceiving a consistent pattern of cooperation metaphors suitable for the different application situations; the designers, in other words, must describe the rules for the collaboration between users.
A generic package for supporting cooperation over the Web, therefore, must not provide a fixed set of cooperation metaphors, but rather must present the designers with a vast range of possibilities that can be selected and tuned to the specific needs of the application. This functionality is lacking in all the Virtual Communities tools available today, and that is why we are designing WebTalk-II, a collaborative environment where powerful cooperation metaphors can be combined and tuned at will.
Cooperation Metaphors for Virtual Museums
With some oversimplification, we may state that collaboration can be represented with two visualization paradigms corresponding to 2D and 3D representations. While both representations are in fact two-dimensional, since the visualization field is a 'flat' computer screen, the main difference between the two is that the first visualization paradigm causes the users to think in term of flat geometric elements and flat structures (windows, icons in windows, buttons laid in a bidimensional grid...), while the 3D paradigm utilizes perspectival projection to convey a sense of spatiality and immersion. Very often this paradigm takes advantage of a subjective view; that is, the users infer position in space by viewing a perspective representation of immediate surroundings, but without the ability to see themselves (exactly as in Real Reality). This is usually the approach taken in gaming and web-oriented virtual reality. In high-end virtual reality systems that make use of complicated devices such as gloves and head-mounted displays, there is an effort to convey much more immersion to the user, by tricking the human eye into seeing depth and space in a synthetic computer generated scene.
Most of the collaboration patterns can be implemented in either 2D or 3D. In 3D, however, it is possible to conceive a set of collaborative situations that are impossible to represent in 2D, and which can convey a very heightened mutual awareness. In the following we will list the most important collaborative elements, and we specify how they could be implemented in 2D or 3D. For some of these cooperation patterns, it is possible to describe situations that have no match in reality, but can be effectively used for the purposes of particular applications. We will point out some of these cases.
At the end of this survey, we will describe sample scenarios with which these collaborative sets can be composed into collaborative behavioural patterns to support a virtual museum application, by using commercial or research tools available today.
Distribution of the Shared State
A shared state is a collection of information about the status of each participant in the application. In 3D environments, it could be the avatar position and its movement or idle state, or the status of a shared resource, like the writing on a collaborative whiteboard. Each user can participate or be isolated from the shared state (and in this last case, it cannot be part of the collaboration mechanisms). The sharing of the state works on two levels:
If space and time sharing are not enabled simultaneously, it is not possible to support collaboration. If space and time sharing are enabled, an Integral or Group sharing mode has to be specified. If Integral sharing is used, it is not possible to enable Group sharing and viceversa.
Enhancement of these collaborative patterns over reality
The possibility of creating applications in which time is counted differently between users, or in which for a same environment there exist several possible states at the same time, has no counterpart in real reality. The first option can be used to let users experiment with artefacts or situations with different speed and timings. The second option is interesting to create multi-path situations in which different choices give rise to different events. An application could model this line of thinking and supply the users with an original way to experiment with this, using Group Sharing.
Grouping mechanisms refer to rules set to the management of user groups. Groups are a means tocluster users by interest or activity; each group shares particular cooperation elements, and behaves following a common pattern. Each group has to be identified by a unique name, so for example a group created to discuss Modern Painting following a renowned critic may be entitled "ModernArtCritique", while a group designed for free browsing around the museum resources can be called "StrollingAroundTheMuseum". Group management is handled by the following four subcategories.
Kicking: other users are allowed to force a user out of the current group. Usually this operation is tied to some kind of privileged status.
A group can be either dynamic or fixed. Similarly, it can be either free or password protected, but both types can have a fixed quota, banning and kicking. Only one type of disbanding can be specified. Leadership management can be defined by using more elements at a time; that is, leadership can be Creation Based and Forbidden, or Creation Based and Tokenized. Some of the management elements cannot be used together, e.g. leadership cannot be Tokenized and Inherited.
Enhancement of these collaborative patterns over reality
Group collaboration in real reality is usually described by a Creation Based, Forbidden leadership pattern, with a fixed/fixed quota group creation and dynamic disbanding. There is no complex notion of leadership management in real groups. In virtual collaboration, leadership can be spread over the group and be used to control enhanced collaboration facilities. (See for example visualization and view-point shifting capabilities).
Information flow rules can be attached to any other cooperation element to specify which are the directions and modalities in which information can be transmitted during collaboration.
Push and Pull modalities can be used simultaneously. For example, a textual message or information item can be pulled or pushed, even if usually the message text is pushed to the user when new messages are incoming (as phone calls are pushed to your cell phone). Push and pull are particularly effective in visualization, where it can be decided if one user wants to see from the point of view (through the eyes) of another user, or conversely if one user wants to force other users to see what he sees (pushing his view onto others). Moreover, it could be possible to push one's position to others, to force other users to determined positions in the virtual space, and so on.
Enhancement of these collaborative patterns over reality
Push and pull are in general common modalities for textual and vocal communication, but are uniquely an achievement of virtuality when it comes to visualization and physical position.
Personal views can all be used, or forbidden for particular application-related reasons. However, only one kind of view at a time is possible. Scene views are all mutual exclusive.
Enhancement of these collaborative patterns
Visualization elements strongly enhance the daily experience of collaboration in reality. In virtual realities it is possible to see oneself acting, or to incarnate other users by seeing from their eyes. This yields enormous gaming, demonstration and collaboration potential.
Specifies which kind of movements are allowed to users in the environment. Movements can be restricted for educational or navigational purposes within the particular application.
Mutual Exclusions: All movement modalities can be allowed simultaneously.
Enhancement of these collaborative patterns
The usual pattern in reality is normal walking in contiguous ranges of time and space. In virtuality the possibility of hyperjumping, not present in reality, enables a more effective collaboration.
Self representation describes the means by which the particular state of the user can be represented. These elements can be enabled or disabled according to their relevance to a particular application.
In most virtual environments, the most effective mean of communication is still text based: users type in what they want to say to other users, and read the answers. Some environments also offer vocal communication, and there is a pattern about governing vocal communication that we will not discuss here. Still, it is hard to operate in a 3D synthetic environment and talk. For a number of reasons, typing is still the favored method at the moment.
Free mode, Group and Group Restricted are in mutual exclusion. The rest of the elements can be used in combination with these first three.
Enhancement of these collaborative patterns
These elements model how talking can happen in real collaboration sessions; however, multiple whisper and group restricted elements may be harder to enforce in reality than they are in virtuality.
Using tools and metaphors for creating collaborative museums
We will now present a short survey of some software tools that enable us to create collaborative virtual environments for museums, commenting on how they can be used to enforce collaboration metaphors in the ways we presented them. Some of these are strictly 3D Virtual Environments (Zyda and Singhal 99), which strive to present a perspectival representation of the environment in which all participants are immersed. Others simply provide users with a way to collaborate over the Internet with usual 2D metaphors. Still, they can be usefully employed for museum applications, and they do use some of the cooperation elements we mentioned earlier.
WebTalk-I (see papers in previous M&W conferences,(Barbieri 2000. T. Barbieri 99)) allows you to draw a 3D environment that represents a virtual exhibition, and publish it on the web. Users that connect to the web page are presented with the 3D representation and can navigate in it together with others, seeing the other users and chatting with them.
There are two co-peration metaphors, fixed and imposed by the framework:
Fig. 1. In WebTalk-I, collaboration is supported via a fixed set of collaboration metaphor-s, like free-formed groups, and a guided tour in which leadership uses a tokenized mechanism.
The goal of the WebTalk-II framework is to provide an environment in which it is possible to design and deploy 3D collaborative environments, acting as a way to improve the access to "traditional" Web sites. There is no assumption or constraint about the technology used to implement the Web sites, nor any implication about the design "philosophy". It is clear, however, that well structured and well behaving Web sites, like the ones that could be obtained using a proper design methodology (HDM and W2000, Rohel, 1997, Bochicchio and Paolini, 1998, Bochicchio and Paolini, 1999, Garzotto, 2001), are targets of specific interest. In addition, there is the further requirement that WebTalk-II must be able to interface to Websites where pages are not statically defined, but dynamically generated (it is in fact much easier to define gateways to a set of statically defined pages, than to a set of pages with number and content dynamically defined), according to user requests. The JWEB (Bochicchio et al 1999) environment is one, of the many available, that does exactly this.
Consider, for example, the need to coordinate access to a number of pages, for a museum application. The user accesses an "index page" that allows "navigation" to the item of interest. In a collaborative environment, the equivalent of the index page is a 3D sub-space, where the users move through visualizing "gateways" to the indexed pages.
In "Virtual Leonardo" for example, the index is represented by a virtual "cloister", while each page is represented by a virtual room. Each room, in turn, contains an exhibit that can be browsed in a collaborative fashion with other participants.
Unlike WebTalk-I where you can have only free groups and pulled visualization from a privileged group user (to form a "guided tour group" metaphor), WebTalk-II allows you to specify (for any virtual exhibition one may design and implement), any possible combination of the collaboration elements discussed earlier in the paper, choosing the best way for people to work with each other given the particular nature of the exhibition the virtual museum is proposing. In other words, you can conceive any cooperation metaphor you like, assign it a name, and specify its properties by enabling or disabling each of the collaboration elements designed above.
Fig. 2. The designer is able to define the general access structure, and specify which rules to enact for the collaborative exploration of the museum (creating cooperation metaphors).
Net2Gether (http://www.net2gether.com ) is an interesting tool for 2D based collaboration over the Web. It provides a chat window under the usual browser window. Chat is organized in groups, and each group has a leader. The leader is able to push the Web page being visited into the browsers of the other participants. It is thus possible to 'share' the 2D navigation of the museum Web site. The tool also allows pre-recording of sequences of chat messages and of navigation events within the site, and plays it remotely to the users. In this way it is possible to create various exploration paths within the exhibition proposed in the museum Web site.
Since Net2Gether has no notion of what cooperation metaphors are, it is not possible to change the way users collaborate. N2G metaphors are thus fixed. They can be classified as:
Microsoft Virtual Worlds
Microsoft Virtual Worlds (Microsoft Virtual Worlds) is an authoring environment for creating 3D Collaborative Virtual Environments (http://www.vworlds.org). Even if there is not an explicit way to model general Cooperation Metaphors in the sense we described in this paper, Virtual Worlds provides the designer with a high degree of flexibility, allowing them to define, for each object, sets of attributes and properties describing them. Some of the most common cooperation metaphors in virtual museums deployed with Virtual Worlds might be:
Fig. 3. An art gallery of a museum can be freely navigated with walking movements in Virtual Worlds (top). Virtual Worlds' Emoticons provide a better way to convey information (bottom).
Conclusions and Future Work
The main point of this paper is that for a large range of applications, cooperation among users while visiting a web site is a very important improvement over current practices. Museum applications belong to this class for a number of reasons:
We have started analyzing the different components for making cooperation effective, recognizing that different situations require different organizations and different actions for cooperation. We have used the name "cooperation metaphors" to identify different cooperation solutions. Also, we have reviewed a number of tools, from the market or our own, in the light of the needs of implementing sophisticated cooperation metaphors.
In the future we will work in the following directions:
Among other application areas, (virtual) museums remain our favorite area of application for these general ideas.
Active Worlds, available at http://www.activeworlds.com
Blaxxun Community, available at http://www.blaxxun.com
Barbieri T., Paolini P. 1999. WebTalk: a 3D Collaborative Environment to Access the Web in Proceedings EUROGRAPHICS '99, Short Papers pg. 111-113, September 99
Barbieri T., Paolini P. 2000. Cooperative Visits to WWW Museum Sites a Year Later: Evaluating the effect, in Proceeding Museums&Web2000, Minneapolis (USA), April 2000
Barbieri, T.1999. Webtalk-II: un'infrastruttura per la cooperazione sul web, Politecnico di Milano, Thesis for the Engineering Degree, Milan, Italy.
Barbieri T. 2000. Networked Virtual Environments for the Web: The WebTalk-I and WebTalk-II Architectures, in Proceedings IEEE for Computer Multimedia & Expo 2000 (ICME), New York, USA, July.
Baresi, L., F. Garzotto, and P. Paolini. 2000. "From Web Sites to Web Applications: New Issues for Conceptual Modeling". In Conceptual Modeling for E-Business and the Web (ER200 Workshops Proceedings), Lecture Notes in Computer Science 1921, Springer.
Benford, S., D. Snowdon, et al. Visualising and Populating the Web: Collaborative Virtual Environments for Browsing, Searching and Inhabiting Workspace, in proceedings JENC8
Bridges, H.A. and Charitos, D., 1997. The architectural design in virtual environments, R. Junge (ed) CAAD Futures'97, Kluwer Academic Publishers, Dordrecht.
Bochicchio M. A., Paiano R., Paolini P., 1999. " JWeb: an Innovative Architecture for Web Applications", in Proceedings of ICSC'99
Bochicchio M.A., Paolini P., 1998. "An HDM Interpreter for On-Line Tutorials", in Proceedings of Multimedia Modeling 1998 (MMM'98), N.Magnenat-Thalmann and D. Thalmann, eds. IEEE Computer Society, Los Alamitos, Ca, USA, pp. 184-190.
Bochicchio M. A., Paiano R., Paolini P., 1999. "JWeb: an HDM Environment for fast development of Web Applications" . In Proceedings of IEEE Multimedia Computing and Systems (ICMCS '99), Vol.2, pp.809-813.
Campbell, D., 1996. Design in Virtual Environments Using Architectural Metaphors: a HIT La Gallery. March Dissertation, University of Washington.
Cerulli, C. 1999. Exploiting the Potential of 3D navigable Virtual Exhibition Spaces, in proceedings Museums and the Web 99, New Orleans
Charitos, D., Rutherford, P., 1997. Ways of aiding navigation within VRML Worlds, Proceedings of the 6th EuropIA Conference, Edinburgh.
Charitos, D., 1996. Defining existential space in virtual environments, Proceedings of Virtual Reality Worlds '96, February '96, IDG Magazines, Stuttgart.
Franca Garzotto, Paolo Paolini, Luciano Baresi, 2001. "Supporting Reusable Web Design with HDM-Edit". in Proceedings IEEE Int. Conf. on System Sciences, Maui (HW, USA), IEEE Press, January 2001-
Gutwin, C., S. Greenberg, 1999. A Framework of Awareness for Small Groups in Shared-Workspace Groupware. Technical Report 99-1, Department of Computer Science, University of Saskatchewan, Canada.
Hearst, M., C. Karadi, 19997. Cat-a-cone: An Interactive Interface for Specifying Searches and Viewing Retrieval Results using a Large Category Hierarchy, proceedings of 20th Annual International ACM/SIGIR Conference, Philadelphia, PA.
ICQ (I-seek-you). Instant Messenger System, http://www.icq.com
Microsoft Virtual Worlds, available at http://www.vworlds.org
Paolini P., Barbieri T., et al. 1999. Visiting a Museum Together: how to share a visit to a virtual world, in Proceedings Museums&Web '99, New Orleans (USA), March 99, pg. 27-35
Rohel, B. et al. 1997. Late Night VRML 2.0 with Java, Ziff Davis.
WebTalk-I Web Site, available at http://webtalk.elet.polimi.it
M. Zyda, S. Singhal, 1999. Networked Virtual Environments, ACM Press.