Making 3D Work for Social Environments
originally presented at the Computer Game Developers Conference, March 1997

Target Audience

This is intended for game designers, producers, programmers, and those who wish to create 3D content that promotes interaction among a large number or community of users.


Advances in both rendering/display technology and communications have made possible the creation of on-line virtual communities. However, just the raw application of these new technologies to the problem is not enough to create a viable social environment. Attention must also be directed at the people who will populate these new on-line communities and the methods available for them to interact within the context provided. It is my desire with this paper and seminar to provide information about those different methods of interaction that will be of help to those designing these new 3D worlds.

First, we will explore why we want to use the virtual environment 3D paradigm to interact socially with others at all. Next, we will look at some of tools for interaction available to the designer of 3D environments. Then, pitfalls of using this new paradigm without thought to the social aspects or interaction methods that are employed by the designer will be examined. Finally, we will explore solutions to the pitfalls listed, and show examples of various interaction methods from the upcoming release of the CyberPark 3D environment that shows how to promote a positive social atmosphere.

Why 3D?

This is the first question asked by many of the current crop of 2D web-centric users and those who find existing 2D chat rooms a satisfactory social experience. It is also a question asked by many who have had poor experiences in some of the earlier attempts at using the 3D paradigm for a social environment.

For large multi-user environments, there is an absolute need for a robust social/information exchange context. Successful multi-user environments provide their members with a rich "tool-kit" to use in interacting with the environment and other members of that environment. Note that this tool-kit does not require complex graphics in order to succeed. Indeed, some of the early MUDs and MOOs succeeded with no graphics at all! It is due, in part, to the diversity and richness of some of these textual contexts that many users feel 3D environments are somewhat lacking. It is a similar problem to that faced by movie directors who try to turn a good book into a successful movie -- they are competing with the ability of the book readers to imagine the world of the book. In the case of virtual environments, in order to match the richness of the textual context, the designer of a 3D world is required to implement in expensive graphics what a textual world can do with just text. So why go to all the trouble and expense of creating a 3D environment? There are several reasons, including the fact that not everyone is a proficient typist, but the most important reason is immersion. Graphical environments, and most especially 3D environments, allow the member to become more deeply intertwined with the world they inhabit.

Some of the abilities offered by the 3D paradigm to immerse the user include:

Given these abilities of the 3D paradigm, we need to explore ways to take advantage of them to promote the maximum amount of social interaction.

Tools for Social Interaction

A successful context for social interaction means offering the users exactly that -- interaction. The more methods available to express that interaction, the more intertwining the context. In a 3D world, we have the advantage of using the real world to show us many of these methods. Some of the these include:

There are many ways to implement these interaction methods. Some of these methods just involve creating a viable user interface, while others, such as game playing or combat, may involve complete design specifications for entire systems of interaction.

For purposes of this discussion, we will divide these interaction methods into the following categories:

For the first case of direct player to player interaction, the 3D paradigm at first appears to be only an indirect contributor to the user interface. In fact, for chatting within CyberPark (as discussed below), we move the bulk of the mechanics for this activity into its own 2D window. However, the 3D world still provides the basis for the user interface. In the case of chatting, this means a player’s location within the 3D world determines the range of casual (non-directed) chats, thereby determining dynamically who receives messages when entered by the user. In essence the 3D world, by its very nature, picks targets for chats based on distance from the user and the quantity of speech that is taking place (the CyberPark chat algorithm shrinks or increases the chat range based on the amount of chat a given user receives). Users do not even think about how this user interface operates since it mirrors a real world activity. This also holds for the dynamic range algorithm: In busy or noisy areas of the world, you can only hear those who are right next to you (think of a party or crowded mall) while in less crowded areas you can hear those farther away (think of a quiet garden setting or an empty field). In CyberPark, this ranging is performed uniquely on each user and stretches from 1 meter to 12 meters, depending on the rate of chat messages being received. The 3D world also provides a selection interface for directing chats to other players and for performing direct actions upon other players. Again, it is a natural extension of a real world activity that users "intuit" without having to have it explained in detail. For CyberPark this is implemented through a right-click menu. Just right-click on the target player and select the desired action to be performed on or with that player.

For the second case of interaction methods – those that involve the environment, the 3D paradigm really excels for creating easily understood user interfaces since again, they often mirror real world actions. This includes actions involving things or articles that may be picked up and carried within the environment – such as picking up and dropping objects, placing objects on surfaces, and the giving and receiving of objects between players. All of these actions can be done by the user by selecting the target location or target of action using the 3D presentation itself to make the selection.

Another example of interactions that use the 3D presentation is that involved with the creation and display of the user persona or avatar itself. By allowing users to personalize their avatar, they are able to express an aspect of themselves or their fantasy selves to others through the 3D presentation. While in CyberPark the actual selection mechanism for body and clothing options are done in a 2D interface (through our "body builder" and "magic mirror" interfaces), the presentation uses the 3D paradigm to show off these selections. In addition, through the use of animation for selected actions, body language can be incorporated into the user’s selections, further utilizing the 3D presentation to fully express an avatar’s individuality. This is done in a primitive way in the current CyberPark implementation by having different walk sequences and chat sounds (embedded sound effects and phrases available in the chat interface) for male and female avatars. For future releases, the mood of an avatar (implemented today only as a user settable value) will also affect the animation, facial expression, and interaction with non-player inhabitants (NPIs) that roam CyberPark.

Finally, the set of complex interactions such as combat and game playing that use the environment as the arbitrator of the interaction can also take advantage of the 3D presentation to more clearly and easily permit these complex interactions. Environmental variables set by the designer such as speed of movement, distance between combatants or between a player and a projectile, range of explosions, etc. are all displayed graphically and noted by the participants. The health of participants, the weapons they carry, the armor they wear, the vehicles they use, etc. can all be noted without resort to gauges, text displays or other non-environmental displays that detract from the user becoming immersed in the gaming environment. This is the principle behind the heads-up display used in the real world for fighter jets, helicopters, and other military vehicles -- keep the pilot or operator from having to look away from the action. I.e., keep them immersed in what is taking place in front of them (and hopefully, by extension, keeping them alive). While these techniques are not directly utilized in the context of a social environment, the principle of immersion for the participants remains the same.

Avoiding Pitfalls of Using the 3D Paradigm for Social Environments

Despite the increased possibilities for creating a compelling world using the 3D paradigm, the designer must be careful to avoid limiting the ability of users to interact within that world. Often the increased graphic complexitiy of a scene requires the designer to decrease the available methods of interaction. For instance, in a 3D world, each additional action permitted to a player requires an animation sequence to implement it. These animation sequences (depending on hardware and software) can be very expensive to implement and/or require a large amount of storage space (for pre-rendered animation). Given a limited budget for animation, the designer finds himself limiting user actions based on the available budget for animation graphics. When confronted with this problem, it is important for the designer to select those animations those permit the greatest interaction for the users of the world. Often, a single sequence can be used for similar actions or actions performed in different contexts (this is especially true for combat scenarios). In extreme cases, some actions that are needed for selected interactions should still be permitted, even without graphic support. Actions that are very quick and which do not affect other users are good candidates for this. Sometimes a sound effect will suffice instead of the animation sequence, again still permitting the action despite a limited budget.

Another potential problem is too much scene complexity. When using a 3D world to promote social interaction, it must be simple for the user to identify and interact with the various other users and environment objects present. This means exploring ways to "unload" a scene of elements that are not required in the 3D context. For example, text chat is inherently a 2D interface. The 3D paradigm does not change the fact that you need to see all the text in order to read a message. Plus, the message needs to stay present and static within the scene long enough to read it. So, why clutter the 3D scene with the 2D text when there are really good 2D methods for displaying text? This is the approach taken in the design of CyberPark, where as many as 100 other avatars at the same time can be present in a scene. A voice bubble appears for 1 second over the head of the player who speaks a phrase in the 3D view ("!", "!!!", and "?" for talking, shouting, and questions). Meanwhile a scrollable chat window is displayed below the 3D view. This permits users to read text at their leisure while keeping the 3D display view clear for more dynamic scene elements. Likewise, the names of players and objects are not displayed in the 3D view unless the user either right-clicks on the player or object, or leaves the mouse hanging over the player or object for 2 seconds (i.e. using a tool-tip interface). The goal here is to keep 2D or textual data off the 3D display except for selected short-term information.

The choice of perspective also plays an important role in developing interaction between players. Many combat games, for instance, prefer a true first person perspective from the user’s eye position. This "real world" view permits game elements such as attacks from behind and sneaking up on a player to be implemented. This is an ego based or "I" perspective… Everything occurs from the point of view of "I". For CyberPark, which is not combat oriented, but geared to create social interaction between its members, the choice of perspective was for a first person removed point of view. The default camera position is up and behind the user. Other players and objects behind the player are visible, as well as players on the other side of crowds or objects (since the camera is raised above the head level). This more cinematic approach is a group based or "WE" perspective. The user is seen as a part of the community instead of separate from the community. It helps to remove the "me and them" mentality that an ego based perspective creates. For CyberPark, the cinematic elements of the camera are emphasized even more by not fixing it to the user. It is free to move along a rubber band path in order to smooth out motion and avoid obstructing structures and objects. Again, this helps to create a more social atmosphere by keeping the camera far enough away from nearby buildings as to make them and any signs or other structure based information easily viewable and selectable.

Another aspect of the choice of perspective is to help in viewing complex scenes with many elements. In CyberPark, the environment supports thousands of users in the same region and will display up to 100 avatars at the same time. Try seeing all 100 displayed avatars using a true first person perspective! By raising the camera up and backing it out of the scene a small distance, all or nearly all of these avatars can be viewed at the same time. Note that another design feature that works in conjunction to make this work is that players occupy space – .5 x .5 meter in CyberPark. This keeps them from getting so close to each other that they cannot be individually selected in most instances.

Does your environment support commerce within the paradigm? If so, you better create a bullet proof monetary system! Even if users are only dealing with fantasy currency, they become rabid if they think they have been cheated or shorted. During the earlier network that INN ran, users threatened to quit when they felt they had lost their "Larrybucks." They would spend many real dollars in pursuit of their fantasy currency! For CyberPark, a "closed" monetary system has been implemented – meaning that money is neither created nor destroyed, but merely transferred between accounts. To create or destroy money requires the manual intervention of a system administrator. A money server was created to support this system that performs and records all monetary transactions. By using this money server as the sole arbiter of the money supply, the integrity of the money supply can be monitored at all times. Member services can also inspect user’s individual accounts and get reports of recent transactions, helping to identify or avoid problems with fantasy currencies reported by members. While this problem is not unique to 3D presentations, it is an important issue and will need to be dealt with by the designer of the environment.

Engineering for Good Behavior

When designing an environment, you as the designers get to create the rules. To make this environment successful, it is important to think about what rules of interaction the environment does and does not allow for your target audience. CyberPark does not permit the use of weapons, touching of players by players, or physical violence by virtue of the fact that there is no option or selection that gives you these actions. However, a future planned adult-themed area called the CyberPunk Zone will offer actions of touching (how much and of what kind is up to the players), use of weapons, and even allow avatars to be killed. We say this area has a "greater depth of interaction" between players…

It is not just the obvious actions that promote a friendly atmosphere, even subtle differences in programming can have an effect on user behavior. For example, when a user chats in CyberPark, the message is sent up to the host and received by all the players who have the speaker in their current chat range. This seems straightforward, but what about the case of the player who just walked away as you finished your message, or the one who just arrived, or worse still, the one just walking by who never even stopped near you? What do these players hear, if anything? To help keep players from appearing rude, we go out of our way to make sure that the location of the start of a move is still in chat range for the 1 second frame during which the player walks out of range. We also checkpoint the user’s progress along a long move operation, with chat messages checked along the way at each bounding checkpoint. Also, when a player first enters a scene via teleport, we actually back up and give to the newly entering player text that may have been in progress during his arrival. All of these steps help to keep players from appearing rude to other players when actions by different players are taken in an asymmetric manner.

In addition, features are designed into the environment to help keep players from appearing rude to others. This includes graphical indications of when a player is away from their keyboard, in a game, watching a game, or on the phone. In all of these cases, a player may not respond right away to a chat directed at them. By graphically indicating their status as "busy" or "out-to-lunch" users will know that the player they are chatting with is pre-occupied elsewhere and not just being rude.

Another prime source of unintentional rudeness is the phone. This has many real-world equivalents that can be carried into the 3D environment if the designer is not very careful. For instance, how often do you find yourself driving behind someone using their celphone? I know I find myself often frustrated as the driver in front chats away without paying attention to the road conditions or other drivers around them. In the 3D environment, unlimited celphone use offers the potential for similar, or even greater abuse. It is a feature that may even by highly desired by the users of an environment, but is it really a feature that is conducive to promoting positive social interaction and hence, a positive experience? This is where the designer needs to be very careful about implementation of a feature that can be easily abused by the user population. In the case of CyberPark, only hosts or sysops have celphones. All other users must go to one of the many phone booths (or game tables) to initiate a call. This does two things: 1) it keeps them out of the main traffic lanes and 2) gives a clear indication that when they are in the vicinity of the phone booth they are probably otherwise engaged. Note that receivers of a phone call do not need to be at a phone booth. For them, we had to add some extra code that moves them away from critical travel paths if they receive a call (i.e. get them away from door entry/exit points, teleport entry points, vending machine access areas, and other critical pathways).

Members of the CyberPark environment are given many ways of expressing acts of friendship towards each other, but very few ways to express hostility. There are many methods for the giving of objects, money, telegrams, or articles to other players, but there are no methods available for the taking of these things from someone else. Currently in the design there is only one object that can be thrown in the park – the tomato, and there is only one target that performing the throwing action will actually work on – the mime (a non-player-inhabitant or AI avatar that roams the park and is often annoying). The idea here is to foster activities that players can participate in together or with each other versus against each other.


With a 3D environment, the designer is creating a sense of "place." That place needs to provide enough activities such that the members of that community have many opportunities for interaction. It needs to balance the activities themselves with the needs of the members to socialize about those activities. For example, in the real world, a group might get together to play basketball for an hour, then go somewhere else to socialize and talk about it for another three hours. The successful 3D environment must provide the tools and place for both of these endeavors to occur.

The 3D paradigm offers wonderful opportunities for the game designer to create rich and robust environments. However, without careful consideration of the type of environment and the needs of those who will inhabit it, these opportunities can be easily wasted or misdirected. A successful environment should contain as large a "tool-kit" as possible (as large as the designer can afford) for players to use in performing interactions with each other and the environment. There need to be activities that players can participate in together that gives meaning to the environment (social activities, games, guest speakers, etc.). These actions and activities should be engineered in such as manner as to foster friendly relationships between player avatars. Since the more subtle real-world signals between players such as body language are not yet present in the virtual 3D world, extra effort needs to be directed at avoiding the perceptions of rudeness that can easily be construed when these signals are missing.

Finally, the designer needs to keep in mind what attracts the members of the 3D environment to come there in the first place – the desire for social interaction and the desire to have fun.

 Back (return to Breighton's Bio)

Home (return to The Dawe Household Home Page)