Museums and the Web 1999

Best of the Web

Archives & Museum Informatics
2008 Murray Ave.,
Suite D
Pittsburgh, PA
15217 USA

Join our Mailing List.

Published: March 1999.


Cooperative Visits for Museum WWW Sites

Thimoty Barbieri , Politecnico di Milano, Italy
Paolo Paolini , Politecnico di Milano, Italy
Francesca Alonzo , Politecnico di Milano, Italy
Giuliano Gaia , Museo Nazionale della Scienza e della Tecnica, Italy
Paolo Loiudice , Politecnico di Milano, Italy

Table of Contents

  1. Introduction and background
  2. Our approach, "Shareable Visits" and VRTalk
  3. The Shareable Visit to the Virtual Museum of Science and Technology of Milan
  4. Conclusions and future work
  5. References


1. Introduction and background    (top)

The notion of "Virtual Visit", to any place, or, more specifically, to a Museum, may take several different meanings; the following could be examples of them:

  1. The virtual visit consists into browsing the web-site of the museum. The organization of the "virtual" content mimics the actual organization of the "real" content within the museum (therefore accepting and reproducing the physical limitation of it).
  2. The virtual visit consists into browsing the web-site of the museum. The organization of the digital content it is not related to the actual organization of the actual content within the museum (therefore neglecting the physical limitation of it, and fully exploiting the possibilities offered by Hypermedia).
  3. The visit consists into browsing within a "virtual" representation of the actual building hosting the museum, with the digital content located and exhibited mimicking the actual exhibition.

Just to make clear what we mean, let us be more explicit:

  1. the "visitor" perceives the content of the museums logically arranged according to actual arrangement within the museum; he/she has little understanding (or no understanding at all) of the museum as a physical facility;
  2. the "visitor" is in "hyperspace", not related at all, nor to the actual logical organization of the content, nor to the physical appearance of the building;
  3. an actual physical visit to real museum is simulated; the "visitor" perceives both the physical structure of the museum, and actual location of the exhibits within it.

Which one is the best way of organizing a virtual visit, cannot be said in general, but it depends upon the goals and the aims of the intended visit.

If the intention is to allow the "visitor" to precisely organize a physical visit to the museum, or to let him/her to recall an actual visit already performed, the virtual visit of type "c" is the most effective. Also, the visit of type "c" is of particular interest when the container, i.e. the museum as a building, is interesting, almost at same level as the content.

If the goal, however, is simply to get the visitor interested in the content of the museum (possibly, but not necessarily, encouraging a visit), the approach b is preferable. Approach "a" falls somehow in between the two extremes, since it communicates the conceptual organization of the physical museum, but it does not attempt to visually reproduce it.

In most cases, in current hypermedia, a mixture of different styles of visits is often used.

Approach "a" and "b" are the most popular, with added, sometimes, a little of "c".

The "classical" CD-ROM [Art Gallery, 1993], for example, uses the approach "b", since the actual organization of the museum is completely neglected. A CD-ROM like [Louvre, 1994] or a site as [] are between "a" and "b," with the latter prevailing: the overall organization of the content is independent from the physical organization of the museum, still the location of the exhibit in each room is also made available,

A CD-ROM such as [Orsay, 1996], provides a good combination of "b" and "c", since, at the same time, the museum can be physically navigated , or the content of the museum can be browsed independently from the physical organization.

For a number of, mostly technical, reasons, approach "c" is not very common today, on the web. If not completely missing is used, in limited amount, as a "divertissement" or as a show of technology. Within [], for example, a limited amount of physical rendering of physical rooms is also available, but it is clearly not being considered as the most important, or a very important way of accessing the information.

The approach "c", in addition has a serious intrinsic drawback: it couples technological problems (the choice of tools and of the delivery platforms, with difficult compatibility among the different choices), with possible low performances (on the web, above all), and with a not to accurate rendering of the visual quality of the exhibits. In all the applications, we have seen so far, the visitor can look, for example, at paintings hanging from the walls, but he/she can’t see them with adequate visual quality such a to appreciate them aesthetically. At most he/she can vaguely recognize the paintings, and often only with the help of the title being shown somewhere.

The main point of this paper is not to discuss merits and demerits of virtual visits nor to discuss technical issues. We wish to discuss, instead, a related problem: how to transform a lonely experience, as are the virtual visits today, in a more engaging experience, where several people can be involved together.

There are computer applications where several people can interact together: forums, chats, virtual communities, etc, are places where several people can "virtually" meet, talk to each other, exchange opinions, share experiences, etc. While engaged in these activities, however, people do not do anything else, but interact each other. It is our goal, instead, to provide the "virtual" visitor of a museum with the possibility of interacting with other people, while being engaged in a immersive experience of visiting a museum, with large and deep information being involved,

A second starting point for observation are the typical "GroupWare" applications Using NetMeeting, combined with digital videoconferences, for example, two o more people can look at each other, talk to each other, share applications (e.g. an hypermedia application or a whiteboard). The interaction has a good bandwidth ) in the sense that it is easy and effective to interact with other people. The virtual visit is missing, however. One of the participants, of course, may start to navigate within a virtual world. The overall setup, however, is rigid (in the sense that the application is actually local to one of the participant) and it is in practice a solitary browsing made visible to others, not a true cooperation.

In the next section we will examine the possible different paradigms of interaction, and how we have decided to implement them in the first prototype.


2. Our approach, "Shareable Visits" and VRTalk   (top)

The starting point is to observe how the people interact each other, while visiting together a museum.

All sorts of different situations may arise, but let us focus, for simplicity, upon three of them:

  1. Visitor X goes alone at the Museum. He/she can browse across the different rooms, focus upon the exhibits of his/her interest, resume the wandering. While he/she is doing this, he/she can look at the other people, see what they are focusing upon, overhear their conversation, listening to the lecture for a group of tourists in a guided tour, etc.
  2. Visitors X, Y, Z … go to the museum together. They talk each other, exchanging comments, opinions, suggestions. Again they can partially share the experience of (some of) the other people being in the museum at the same time. Sometimes X takes the lead, conducts the group, and guides the conversation, etc., with the other members following him/her, listening to him/her and also exchanging opinions among themselves. At other times Y (or Z..) takes the leading. The group, anyway, is a loose one: its members can always go by themselves, or they can stop somewhere, or form subgroups, …. At moments, however, the members of the original group rejoin together and proceed in the visit.
  3. There is an organized tour: a group leader goes across the museum, pointing at the exhibits of special interest; stopping whenever is needed, providing comments and explanations. Most of the times the members of the group closely follow the guide and listen to him/her. Sometimes, however, they "whisper" comments each other, or they stop somewhere, in order to look to something of their interest, or they wander freely around, and then go back to join the group. After a while, however, the memebers join again the group and continue following the guide.

A countless number of variations and combination of the above-mentioned paradigms, i.e. ways of visiting a museum can be devised. We focus, for simplicity, upon them calling them, respectively, "free-visit", "loose-group" and "guided-group".

In the rest of this section we will examine, briefly, what type of functionality we would need in order to support those types of visits, while in the next section we will examine our specific technical approach, and a prototype application being developed in cooperation with the Museum of Science and Technology of Milan.

Our goal is to allow the user experiencing a "virtual visit" to share, at some degree, the visit with other users. Our starting point has been an analysis of the three paradigms above described, end trying to implement them in a virtual visit.

Assuming the subjective point of view of a virtual visitor, the following elements of a virtual visit have been analyzed:

  2. Whom other visitors may I see.

  3. SEEN
  4. Who can see me.

  5. SCENE
  6. I could look at the scene from fixed points of view, or from my eyes, from the eyes of someone else (e.g. my friend or my group leader), …

  8. My movement within the virtual scene can be free, can follows predefined paths, can be dependent upon the movement of somebody else (e.g. my friend or my group leader), ….

  9. TALK
  10. I can talk to everybody else ("broadcast"), I can talk to a single person ("whisper"), I can talk to a small group of people ("chat"), …

  12. I can listen to everybody, I can listen to a single person, I can listen to a small group, ..

  14. The messages can remain the same everywhere within a virtual world, the messages can degrade with distance, the messages can be confined within a world, or span over to neighboring worlds, ….


The objects within the virtual world can be "duplicated" (each one gets his/her copy), shared without possibility of manipulating them, shared with possibility for everybody of manipulating them, shared within a group but duplicated across groups, …

Several other elements could have been considered, of course, but the above list seemed long enough to us for an initial experiment.

The question then naturally arises: what type of technology should be used for the virtual visit?

Without getting into technical details, for the time being not interesting, we rapidly came to the following conclusion: 2D techniques (the standard Web techniques) are very good at providing information (images, text, graphic), and very bed for grasping the overall aspect of a complex situation, with a complex arrangement of objects, and several people sharing the situation. 3D techniques (either photographic or graphic) are very good for grasping complex environments, with several people, and quite bad at communicating information (images, text, video, etc.) of acceptable quality.

Since our goal was to experiment with cooperation, while performing virtual visits, we have chosen 3D virtual environment as the fastest way to get an experimental prototype. Therefore in our current approach 3D is used to browse the museum and to find the way around, while standard WWW techniques are used to provide more substantial information.

VRTalk is our experimental environment; it is a powerful tool that allows creating Virtual Reality three-dimensional worlds, in which people can meet, by means of an Internet connection. Each one of the virtual visitors can remotely explore the Virtual world. In addition a virtual visitor can examine and interact with components of the world. Like in a real world situation, a visitor can see where the other visitors are currently located, where they are going, and what they are doing. Visitors exchange opinions or information, with other visitors, through the keyboard. Visitors can interact with 3D objects, sharing the experience of the interaction, with other visitors.

At the moment the virtual world must be a VRML world (a 3D-Java version is planned for summer 99).

With respect to standard VRML worlds a few extra lines in the code are needed in order to declare which objects are "shareable" with other users. Our experience is that any "reasonable" VRML world can be shared, at a substantial degree. The little modifications required today, need an experienced VRML programmer; by the summer 99 a tool is envisioned in order to automate the transformation.

In terms of delivery, any user connected via Internet, that can execute a VRML application (i.e. that has the proper VRML plug-ins installed) can access a VRTalk application.

In terms of performances, we have experienced excellent results using Intranets based on LAN’s. Using standard Internet connections degrades performances, of course; in some cases this degradation can be very disturbing, spoiling the user satisfaction. In other cases, however, low performances can be even helpful, since in a cooperative visit everybody should move and react slowly, in order to get everybody else understanding what he/she is doing or attempting to do. This point will be exemplified in the next section.

Explaining how VRTalk works, without going into technical details, it must be said that the shared application runs on a server. All the clients, i.e. the machines of the actual users, have a local copy of the shared application. When a client machine (i.e. a visitor) "acts" upon the world, it "sends" a notification of they action to the server; the server collects all the notifications of actions, and sends them back to all the clients. Each client, in turn, when gets notified of an action performed by someone else, updates its local copy of the shared application. The mechanism above described explains why there can be a certain delay between the action performed by Visitor X, and the awareness of the action by Visitor Y. Therefore the application is conceived for "well behaving" visitors; in other world VRTalk supports a cooperative application (where everybody tries to help everybody else), and not a competitive application, where someone tries to compete with someone else for resources.

The user of a VRTalk application sees a browser window, which is split in two parts. The upper half shows the 3D representation on the world, in which the visitor can move and interact with objects. In this part the user can also see other visitors moving and performing actions. A human-shaped figurine, named avatar, represents every visitor. In the lower half a chat window is provided, in which visitors can write messages to other visitors, and can read incoming messages.

Textual, audio and video information can be linked directly within the three-dimensional world, or by popping up contextual 2D web pages along the VRTalk browser window.

In terms of shaping the cooperation pattern, at discussed in the previous section, at the moment we are experimenting with very simplified features, while a more sophisticated environment is envisioned by the summer 99. At the moment visitors can either belong to the "global environment" or can cluster themselves into Groups. The cooperation rules are the following:

  • Global environment
  • A visitor sees the world as populated with the avatars of the other visitors (placed according to the position of each visitor).
  • A visitor can see through "his eyes" (subjective camera) or through a set of standard cameras; in the latter case he/she can’t see his/her own avatar.
  • A visitor is aware of whoever is in the world (through an explicit list of names).
  • The chat is global in the sense that everybody can speak and everybody gets what is being said.
  • Groups
  • Each visitor can play the role of group leader, but only one visitor can play the role at any given time point.
  • A visitor sees the world as populated with the avatars of the other visitors (placed according to the position of each visitor), of his/her own group only.
  • A visitor can see through "his eyes" (subjective camera) or to a set of standard cameras; in the latter case he/she can’t see his/her own avatar.
  • A visitor can also see through the eyes of the group leader, i.e. can see at what he/she is looking at.
  • In the normal situation the visitor can move around independently from the group leader, and it is up to visitor to try to "follows" the leader.
  • If, however, the visitor sees through the eyes of the group leader, he/she is physically tied to the group leader itself, following all its movements. When the visitor drops off the connection with the group leader, he/she will find himself/herself in the physical position of the group leader.
  • A visitor is aware of whoever is in the world (through an explicit list of names), but only within the limits of the group.
  • The chat is global only within the limits of the group.

In the next section we briefly illustrate, mainly through pictures, a prototype application, which is a virtual tour within a portion of the MST (Museum of Science and Technology) of Milan. It is clear that it is impossible to reproduce on paper the actual "feeling" of a cooperative virtual visits, but only a few hints are provided.


3. The Shareable Visit to the Virtual Museum of Science and Technology of Milan   (top)

The prototype application allows a virtual visit across a few rooms of the MST. The MST hosts several objects, artifacts, pictures and documents concerning the development of science and technology around the world, but with a specific emphasis upon Italy. The museum is hosted in an old building, which was used once as a monastery; one of the nicest spot, in fact, is a cloister. Among the objects hosted by the museum, specific relevance have the "Leonardo machines": they are wood machines (build in the period across the end of the last century and the beginning of this century), trying to physically recreate what is understandable from the drawings of Leonardo da Vinci. Most of the machines do actually work, in the sense that the visitor of the "real" museum may operate them.

In our virtual visit we have reproduced the cloister and a few rooms; we have "slightly" modified the actual floor plan, since the true one was a little "boring" for a virtual visit. The rooms are "populated" with a few interactive virtual machines, in the sense that that the virtual visitor may interact with them. A few billboards are used to access standard Web pages providing explanatory text.

Therefore, overall, our virtual reproduction is quite crude or poor. Our goal, at the moment, is mainly to gain experience with cooperative visits; by summer ’99 a more refined representation of the museum is planned.

In the experiment, used to generate the pictures, there were four virtual visitors.

Figure 1 shows the point of view of one of the visitors: he/she sees 2 avatars representing two other people currently visiting the cloister.


Figure 2 shows the cloister with three other visitors and a chat (lower half of the page) going on among them.


Figure 3 shows a room with one of the machines (a giant screw) that is possible to interact with. The visitor is actually looking at a group leader.


Figure 4 shows the same situation depicted in picture 3, with the visitor looking through the eyes of its leader, i.e. looking at what the leader was actually looking.



4. Conclusions and future work   (top)

The authors of this paper have the firm belief that transforming virtual visits, from an (essentially) lonely experience, into an experience that can be shared with other people is a very interesting goal. Applications can range from commercial ones (cooperative shopping) to cultural ones. Sharing with friends a virtual visit to a museum, for example, or being lectured by an expert while virtually visiting a museum, can greatly enhance the pleasure and the cultural outcome of the experience.

Several problems still make the goal not easily attainable, and we discuss below a few of them:

  • visiting a museum while interacting with other people

Interacting with other people, while virtually visiting can be helpful, or can be disturbing; the same medium (a very small screen) is used for both activities and the overall "bandwidth", i.e. the visual communication capacity, is quite small. In addition current virtual museums assume that the full screen is available to them; subtracting part of the small screen for other purposes can spoil the whole experience. So, should we design virtual visits (either 2D or 3d) with specific features, in order to make them suitable for cooperative visits? On the other hand it is appealing to allow any virtual visit, as currently designed and implemented, to become shareable with other visitors

  • cooperative visits metaphors

As we have briefly discussed in the previous sections, there are several ways to organize the group of visitors who can cooperate, several ways to organize what each visitor can see or can listen to, several ways to coordinate the movements around, etc.

Our opinion is that we must consider real-life interactions among visitors as a starting point for requirements, in order to build the initial prototypes. With these prototypes we can learn about what virtual visitors like to use or do not like to use or would desire to have. Later, on the basis of the previous experiences and user reactions, we can develop better metaphors and paradigms.

  • 3D and 2D integration

As we have said in previous sections, we are convinced that 3D graphic is very god at conveying the physical appearance of a building and the presence (or absence) of objects and people in it. 3D technology, however, at the moment and for a while, does not have the capability of communicating high quality content.

2D technology, on the contrary, can easily deliver high quality images, text and graphics; it does not easily allow, however, to recreate the feeling of being somewhere, and, even less, of being with someone else.

One possible development is to find a better way to combine 3D graphics, with standard ways of organizing web-sites. We need better logical organization, and smoother ways to perform the transition from one representation to the other one.

Another possibility is to improve the quality of the 3D representation, bringing it at the same level as the standard 2D; this possibility, however, is tied to the development of technology (both HW and SW), rather than to the to ability of the application developers.

  • performances

Performances are an issue for cooperating virtual visits; it takes a while, over Internet, in order to notify every virtual visitors of what the other visitors are doing. This inherent slowness can create three types of problems:

  • Annoyance of the visitor, for the slowness of the response.
  • Application confusion and difficulty of coordinating the interaction.

Think, for example, of visitors X and Y, both wishing to move an object a little bit on the right.

X does it, and after a while Y (who has not received, yet, the notification of the action of X) does the same; the result is that the object gets moved twice as much as it was desired. If now both visitors want to remove the anomaly and move the object a little on the left, the same wrong procedure can be repeated.

  • System trash.

The slowness of the incoming messages and the sudden bursts of messages coming from several clients at the same time may confuse the server.

On the ground of the above considerations, the following is our research agenda for the near future:

  • to better understand the nature of a cooperative virtual visit, and to improve (also completing) the prototype application for the MST, in order to meet those requirements.
  • To define a wide range of ways of cooperating for virtual visitors and to implement them.
  • To better coordinate the relationship between the 3D environment and a standard, well-structured, WWW site, also improving the transition from one representation to the other one.
  • To develop a better communication among visitors, in order to improve their interaction, and also trying to avoid chaotic cooperation.

The above agenda should be completed by the summer ’99. Even before that time (around the spring 99) we will start, in cooperation with the Museum of Science and Technology, collecting data about the user satisfaction and about usability of the application. The data collected from usability testing and the user reactions, will be the ground to devise better ways to let the "visitors" to cooperate, while performing their virtual visits.



5. References    (top)

Rohel B., Couch J., Reed-Ballereich C., Rohaly T., Brown G. (1997). Late Night VRML 2.0 with Java. Emeryville (CA): Ziff-Davis Press.

Journal Articles:
Garzotto F., Paolini P., Schwabe D. (1993). HDM - A Model Based Approach to Hypertext Application Design  in ACM Trans. on Information Systems, Vol. 11, No. 1, Jan.

Garzotto F., Mainetti L., Paolini P. (Oct. 1996). Navigation Patterns in Museum Hypermedia in Proc. ICHIM'95 - International Conference on Hypermedia for Museums, S. Diego (CA).

Garzotto F., Mainetti L., Paolini P. (Oct. 1996 ).  Navigation Patterns in Museum Hypermedia in Proc. ICHIM'95 - International Conference on Hypermedia for Museums, S. Diego (CA).

Garzotto F., Matera M., Paolini P. (April 1998). To Use or Not to Use? Evaluating Usability of Museum Web Sites  Proc. of M&W'98 - International Conference on Museums and Web, Toronto, Canada.

Electronic Materials with No Printed Analogue:

Reunion des Muséés Nationaux (1994). Le Louvre. CD ROM. Mountparnasse Multimédie.

Reunion des Muséés Nationaux (1996). Musées d'Orsay. CD ROM. Mountparnasse Multimédie.

Art Gallery (1993) The Collection of the National Art Gallery, London. Microsoft.

Web sites:

Lombard University Consortium Internet Access:

National Gallery of Art:

Silicon Graphics - References to Cosmo Software Products