there a way to set up a chess game where a 6-year old can match up a
grand master? Can a blind quadriplegic compete in a massive multiplayer
game against sighted, non-motor-disabled, gamers? Can two people
cooperatively share a role playing game, when one of them is using a
mobile phone and the other a next generation game console? These represent just a few examples of the type of questions that the Theory of Parallel Game Universes aims to address.
At this point, it should be clarified that in the context of this article, the term “accessibility”, is not associated only to people with physical, sensory or mental disabilities (which is typically the case), but it also refers to all those gamers that may not be able to fully experience - or even play - a game due to:
- the environment they operate in, e.g., a person in a noisy environment is situationally deaf, someone using a screen in bright sunlight is situationally blind, and so on;
- the hardware and software they use, e.g., a mobile device with a small screen, an older browser, a different operating system, etc.;
- their gaming skills and preferences, e.g., a person who does not like or, experiences difficulty in using 10 different keys in order to kick the ball in a football game.
As a means to achieve game accessibility that is coupled with high interaction quality, the Human Computer Interaction Laboratory of ICS-FORTH1 has introduced the notion of Universally Accessible Games2,3 (UA-Games), as games that:
- follow the principles of Design for All4, being proactively designed to optimally fit and adapt to different individual gamer characteristics without the need of further adjustments or developments;
- can be concurrently played among people with different abilities, ideally also when sharing the same computer;
- can be played through alternative technological platforms and in alternative contexts of use, using a large variety of devices, including assistive technology add-ons.
In other words, UA-Games possess the ability to adapt their interface and content to best serve the requirements of a specific gamer under specific gaming conditions. For example, a Space Invaders type game (Figure 1a), would use enlarged, simplified, graphics when played by someone with deteriorated vision or someone using a small screen (Figure 1b), and would decrease its speed / complexity and employ 3D sound when played by a blind person or by someone who is not using a screen at all (Figure 1c).
To illustrate the concept and prove the feasibility of UA-Games, two fully-functional games have been developed, which are both freely available on the Web: UA-Chess5, a two-player web-based chess game, and Access Invaders6, a multiplayer remake of Space Invaders.
The underlying vision of UA-Games is that through such games people will be able to have fun and compete on an equal basis, while interacting easily and effectively, irrespective of their individual characteristics, the deployed technologies, or the location of use.
At this point, it should be clarified that when referring to games that are universally accessible, it is meant that these games can be played by all people who can potentially play them but may currently be restrained from it due to game design flaws, not by everyone in the world. It is obvious that there will always be games that, due to their intrinsic characteristics, cannot be made accessible to a range of people (e.g., complex strategy games for the cognitive disabled), or when made accessible may have no meaning or interest for those people (e.g., a “find the song title from listening to a melody” game for a deaf).
Introducing the Concept of Parallel Game Universes
A stumbling obstacle encountered while developing Access Invaders was how to support multiplayer sessions where people with diverse “disabilities” could play the game cooperatively, being fully aware of each other, while at the same experiencing the game in an optimally adapted way.
The concept of Parallel Game Universes (or, in short, PGUs) was conceived as a solution to this problem. The suggested approach is to allow each player to play in a different “game universe” and then somehow project each universe to the other(s). A “game universe”7 is defined as an instance of the game after it has been adapted to best suit the requirements and needs of a particular gamer playing under particular conditions. For instance, the examples presented in Figure 1 all represent different game universes.
In order to further illustrate the basic concept, consider the following situation. Two friends want to play the game together. One of them (Player X), due to severe motor-impairments, can use only a single switch. To be able to play the game, her spaceship should be automatically moving and firing, while the player’s interaction is limited to altering the direction of movement by pressing one switch. Due to the auto fire option, the player’s bullets should not collide with the shields, so that they are not accidentally damaged. Furthermore, to achieve an appropriate difficulty level, only a small group of aliens should be introduced that moves slowly and fires very scarcely (see Figure 2a). The second Player (Y) does not have a physical impairment. In order to find the game challenging enough, she prefers to confront numerous fast, fire blazing aliens (see Figure 2a). If the two players attempt to share the very same game, in case it is adapted to player X, then player Y will find it rather boring and would also be able to easily destroy a lot more aliens than X, while if the game is adapted to Y then it will probably be extremely difficult - if not impossible - for X.
Following the idea of Parallel Game Universes, a possible solution is to merge the two distinct game universes into one (Figure 3). Thus, in this new game, two groups of aliens would exist: a big, fast and powerful group which can destroy and be destroyed only by Player Y, and a small, slow and quite harmless group that plays only against Player X. The bullets of each player would not affect the aliens “belonging” to the other player, while Player X’s bullets would not collide with the shields, and Player Y’s would.
Figure 3: Example of how the concept of Parallel Game Universes can be applied.
Generalizing the Concept
In the aforementioned example, the process of merging the two distinct game universes into one is quite straightforward, since they both have compatible rendering requirements. A new problem arises when two (or more) universes have competing rendering needs, e.g., a game universe of a person with deteriorated vision where few, large sprites are presented, versus a game universe of a fully-sighted player, where numerous small sprites exist, or a universe of a blind versus the one of a sighted player, where conflicting requirements for aural output are imposed.
In such cases, the concept of Parallel Game Universes can still be implemented, but additional computational resources are required, such as extra sound cards and earphones for providing dedicated sound output to each player, or multiple computers for projecting different views. Furthermore, a “transition function” is needed for translating the events of one universe to the others in a format that is suitable and meaningful for these universes. It is important to note that the overall objective is not to recreate everything that exists or happens in a universe to every other universe, but just to communicate enough cues, so that the players can cooperate in a successful and enjoyable manner. For example, it is desirable for a blind player to know that her sighted game partner has still a few or several aliens to destroy in order for a level to be successfully completed.
The Case of Competitive Multiplayer Gaming
The concept of Parallel Game Universes does not apply only in the case of cooperative games, but also when players are competing. In this case, a key accessibility problem is how to make the game fair, since the players’ skills may be completely different. In other words, players’ weaknesses need to be compensated for. A plausible solution is to delegate part of the game control to a “third” party, which can be either another player, or the computer, thus leading to two alternative options:
- a. Collaborative gaming. Two (or more) players act as one. Game control is shared among the players much like in the way it was in World-War II fighting biplanes, where one person handled navigation, while another was responsible for shooting. An illustrative example of a hardware device that supported this gaming paradigm was Team Xtreme8 by Pathways Development Group, Inc. Team Xtreme was a hardware box for N64 in which 1 to 5 switches could be plugged to control any keys of the game controller. This device allowed a player with disabilities to team-up with another person, who assisted using a standard game controller.
- b. AI-supported gaming. Typically, artificial intelligence (AI) is used in computer games to control the non-player characters (NPCs), such as the monsters in a platform game, the “bad guys” in a fighting game, the opponent brain in a strategy game, the player’s sidekicks in a role playing game, etc. However, it can also be used to compensate for individual player weaknesses (e.g., novice vs. experienced player, single-switch gaming vs. full game controller), and work with the player in a synergetic way, similarly to the way another human would in the case of collaborative gaming. Thus, AI-supported gaming has the potential to allow a player to compete on an equal basis against the computer, or any other player, irrespective of individual (dis)abilities.
Figure 4: Example of 2 players competing in Parallel Game Universes.
As a related example (see Figure 4), consider a tennis computer game where a person who is hand-motor impaired (Player A) plays against a person with no hand-motor impairments (Player B). In the universe of Player A, the game is rendered as a simple 2D game that resembles the game of Pong. The player must place the bat underneath the ball using 2 switches, for moving up and down. In Player’s B universe, the game is represented as a 3D realistic tennis simulation seen from a first-person viewpoint. The player controls an accurately rendered model of an athlete using a gamepad’s joystick and 8 buttons. In order to hit the ball, player B has to position the athlete correctly in space, and also adjust the height and movement of the athlete’s arm. The two universes are “synchronized” through a transition function (ƒT) which is responsible for translating the ball’s and players’ 3D positions and speed vectors to 2D, and vice versa.
The only limit to what can be achieved through PGUs is the design of the transition functions used. A rather challenging example is presented in Figure 5, where two sample transition functions (fT and gT) are presented for making a match of chess accessible to people with different mental skills. In this case, the transition functions use AI to transform the complex problem of selecting the best possible move from the chessboard to a much simpler one. Of course, since the simplified versions of the game may take less moves to finish, a single match in one universe (e.g., the full chessboard) may correspond to several matches in another one (e.g., tic-tac-toe).
It is obvious that in this example, the mapping between the two universes is not straightforward, due to the significant differences among them. Some readers may even think that it is not possible to find an adequate transition function, since in the case of chess the possible moves’ space is subject to combinatorial explosion, while in the case of tic-tac-toe it is restricted and manageable. Still, according to the theory of PGUs, the designer’s goal is not to find a computationally equivalent model among the two universes, but to devise a creative solution so that two players with highly diverse skills can compete on an equal basis.
An interesting real-life example of Parallel Game Universes took place in August 18, 2000 on the Pelican Hill golf course in Newport Beach, California, where people with quadriplegia and paraplegia played golf side-by-side with able-bodied players9. The only difference was that the people with disabilities made their shots virtually, using a wheelchair-mounted computer equipped with Madentec’s assistive technology and Microsoft's Links golf software with a detailed model of the course, and then they were following the path of their virtual ball on the actual course.
At this point it should have become clear that PGUs are not just different game “skins” or views, since it is not only the game appearance and viewpoint that changes but also its content, or even the game logic.
Figure 5: Example of two potential transition functions for making a match of chess accessible to people with different mental skills.
Key Properties of Parallel Game Universes
PGUs are characterised by two key properties: individualization (Figure 6) and balance (Figure 7).
Figure 6: PGUs support individualization by extracting the pure essence of games, which is game experience, and offering it to the individual players according to their needs and preferences.
Figure 7: PGUs strive for balance by compensating for individual player weaknesses and challenging player strengths, ensuring that opposing forces (player vs. player, or player vs. computer) are matched.
The Four Fundamental Laws of Parallel Game Universes
Parallel Game Universes are governed by the following 4 fundamental laws:
Law #1: A PGU should always adapt itself (i.e., its user interface, content and rules) to best serve the needs and preferences of the active player and the characteristics of the current context of use.
of use” entails the type of the game (e.g., action, puzzle, RPG) and
its particular characteristics (e.g., cooperative, competitive,
content, rules), the deployed technologies (hardware and software), the
physical environment (e.g., indoors, outdoors, home, office), as well
as its condition (e.g., lighting, noise, space).
There are three possible ways for a PGU to adapt:
- Pre-game accessibility adaptation: Before the game starts, the player creates a personal profile stating individual interaction preferences and requirements, as well as information about the current context of use that cannot be automatically inferred by the system, so that when the game starts, adequate means for enabling accessible interaction are provided. This type of adaptation is of paramount importance for achieving a minimum level of accessibility, since for example there is no way for a game to infer that the player is blind, or that she is located in an outdoor environment with bright ambient lighting.
- Pre-game player profiling: In order to achieve game balance (see section 5 above), beyond ensuring a basic level of accessibility, it is also required that game attributes such as difficulty, speed, layout, content, etc. are adapted to the current player. The easiest way to do this is to directly ask the player (or a peer, e.g., teacher, relative, friend) to subjectively select what she considers to be “adequate” values for these attributes. A more challenging, and promising, approach is to use an interactive profiling application (e.g., like a mini-game) that first tests and assesses the player’s abilities that are required by the game, and then automatically generates the required attribute values.
- In-game player monitoring and dynamic adaptation: The player’s actions are monitored during gameplay, and related inferences are made. These inferences are subsequently used to dynamically adjust the gameplay to better match the player’s skills.
Law #2: Each distinct PGU is ruled by its own laws. Any game element, no matter if user- or computer-controlled, that enters the PGU must conform to these laws.
Law #3: Regarding any two PGUs (A and B) a game element can be in one of the following states: private, shared, or, monitored.
- Private: The element exists in just one of the PGUs and can not affect or be perceived in other PGUs; e.g., a set of protective shields may be present only in universe A, while their existence is unknown to universe B.
- Shared: The element exists in (at least) two PGUs. In this case, Law#2 should be applied for rendering the element in each universe. Thus, for example, an element generically known as “vicious alien” may be rendered as an ugly, ferocious, monster in universe A, and as a funny, goofball, cartoon character in universe B. Players can also concurrently exist in more than one PGUs (e.g., through split screens or multiple projections). In this case, these alternative PGUs must be mutually “compatible” and accessible to the specific player. If, for any reason, a shared element is destroyed in a PGU, then it must be automatically destroyed in the other. This state is characterized by a principle that is also referred to as “loose consistency”.
- Monitored: The element exists in and can affect only one of the universes, but can also be perceived in the other, but without any effect. For example, in universe A, a blind player hears very loud and clear the sound of the single alien she is competing against, but in the distance she can hear the sounds of a battle where another player is fighting in universe B against a horde of aliens.
Law #4: The state of any interactive element can dynamically change at anytime (by its own will or by force), as well as the PGU it is located in.
Thus, for example, if a player has destroyed all the aliens in her PGU, she can lend a helping hand to a player in another PGU by asking to send over some more aliens. In this case, the “transferred” aliens can either move or be shared between the two universes. Of course, the same action could also be initiated by the aliens. Thus, if some aliens are having a hard time in a PGU, they may decide to ask for reinforcements from other PGUs, but only if this does not break Law #1. Furthermore, a player may also decide to leave one PGU for another, for several reasons, including escaping from an inevitable fatal situation, or playing the game from a different perspective when stuck in order to find an alternative solution.
Towards a paradigm shift in multiplayer gaming and game accessibility
The potential impact of the PGUs theory is five-fold, since it:
- a. opens up and enhances an entertaining social experience that would otherwise be unavailable to a significant percentage of people;
- b. allows for social interaction among people who may never have (or even could have) interacted with each other;
- c. provides a novel approach for developing massively cross-platform multiplayer games;
- d. considerably expands the life-cycle of games, as well as the size and composition of their potential market;
- e. creates a new model for producing games on different platforms, as well as accessible games, since a company may sell the rights to, or subcontract, a specialized group for developing PGUs for specific platforms or user groups.
The concept of PGUs and the way it is applied can only be limited by the imagination of the game’s designer. It can provide unlimited, unconstrained fun if appropriately used, but it can also be quite disturbing for some game creators. One can easily imagine a small bunch of puzzled game designers asking themselves: “but, if I have to design a game that anybody can take and transform into something completely different (over which I have no control), then what am I supposed to design anyway?”
A possible answer to that question is simply “just exciting game experiences, which is what you should be designing in the first place anyway” :-)
Parallel Game Universes have the potential to allow people with extremely different characteristics and abilities to play together, or even against each other, on an equal basis. Of course, someone may claim that those people will not actually be playing the very same game. It may be so, but does it really matter if one player fights against one slow, naive alien while another struggles against a horde of merciless galactic villains? Or, in case they play against each other, if one of them needs the support of the computer’s AI in order to adequately compete?
The answer to those questions may be highly subjective, but in my opinion, what ultimately matters is that these people are given a chance to share the maximum fun and challenge that they can get from the game, without having to compromise or sacrifice their personal gaming experience due to their individual differences.
Thus, in a broader scope, the theory of Parallel Game Universes is not just about games. It is actually a policy statement. PGUs support and promote two fundamental human rights: individuality and equality. And on top of that, they allow people to experience the (game) world from alternative, novel viewpoints, broadening their perspectives and opening up opportunities for new, fascinating, experiences.
Access Invaders supports the instantiation of Parallel Game Universes on a single computer, following the approach illustrated in Figure 3. You can download a demo from: http://www.ics.forth.gr/hci/ua-games/access-invaders/download.php
How to play:
- 1. Use the mouse or, the up/down arrow keys and ‘space’ or ‘enter’ to select ‘Parallel Game Universes’.
- 2. Press any key twice.
- 3. Select ‘New Game’.
- 4. Player 1 (blue spaceship) uses the left/arrows to move and ‘space’ to fire. Her rockets are quite slow and collide with the shields. Only one of her rockets can be active at any time. She can destroy only the group of insect-like aliens, but she can be destroyed by both groups. The insect-like aliens are fast and fire constantly, while the robot-like aliens are slow and fire at a very slow rate.
- 5. Player 2 (yellowish spaceship) uses the ‘tab’ key to switch its direction of movement. The spaceship is constantly moving and firing. Her rockets are fast and do not collide with the shields. Up to three of her rockets can be active at any time. She can destroy and be destroyed only by the group of robot-like aliens.
I would like to thank Professor Constantine Stephanidis and Dr. Anthony Savidis for their invaluable advice and support.
- Human-Computer Interaction Laboratory, of the Institute of Computer Science, Foundation for Research & Technology – Hellas, http://www.ics.forth.gr/hci/
- Grammenos, D., Savidis, A., Stephanidis C. (2005). UA-Chess: A Universally Accessible Board Game. In Proceedings of the 3rd International Conference on Universal Access in Human-Computer Interaction. G. Salvendy (ed.). Las Vegas, Nevada, USA, July 2005. Lawrence Erlbaum
- Grammenos, D., Savidis, A., Georgalis, Y., & Stephanidis, C. (2006). Access Invaders: Developing a Universally Accessible Action Game. In K. Miesenberger, J. Klaus, W. Zagler, & A. Karshmer (Eds.), Computers Helping People with Special Needs, Proceedings of the 10th International Conference,ICCHP 2006, Linz, Austria, 12 – 14 July (pp. 388–395). Berlin Heidelberg, Germany: Springer. URL: http://dx.doi.org/10.1007/11788713_58
Dr. Dimitris Grammenos
HCI Laboratory, ICS-FORTH
Heraklion, Crete, GR - 70013 Greece
tel: +30 2810 391755, fax: +30 2810 391740
e-mail: [email protected]