JCGD Volume #1 Issue 1 April 1987
One useful way to view the nature of computer games is through the concept of the "interaction circuit". This is the flow of information between the player and the computer in the course of a game. Consider first the flow of information between artist and audience in any other artistic medium:
Artist --> Audience
Artist talks, audience listens. It’s a one-way flow of information, presumably justified by the vast intellectual and moral superiority of the artist. With a computer game, though, we have something entirely new and, I think, much superior:
Computer talks <-- Computer processes <-- Human thinks <-- Human talks
I call this little diagram "the interaction circuit", and it illustrates the nature of the relationship between the game and the player. A closer examination of this circuit suggests some guidelines for good game design. I believe that the quality of the game-experience is in some way monotonically related to the amount of information flowing around this circuit. In other words, the more stuff that moves around this circuit, the more intense or interesting the game will be to the player.
Now, how do we achieve this goal of maximal information flow through the interaction circuit? The trick, I think, is to maximize the performance of each of the four nodes in the loop. We want to design each of the four nodes to offer maximum horsepower for driving lots of information around in the loop. We also want to insure that no single node acts as a bottleneck in the information flow. I shall examine each of the nodes in turn.
I begin with the topmost node, "computer talks". This is primarily what happens on the screen, with the possibility of additional information transfer through the medium of sound. This is the easiest node for computer game designers to understand, because it is most like the old-fangled means of imparting information ("I talk, you listen"). For this reason and others, this node is the one that gets most of the attention from the designer and most of the resources of the computer. Some games devote almost all of their total resources to this single node. It’s a real shame, too, for this is the most uncomputerlike aspect of computer game design.
Why do people bother making computer games that are really little more than movies or novels, especially when they’re such lousy movies or novels? An old military maxim is "Fight on the ground of your own choosing", meaning that one should fight only under the circumstances best suited to the particular strengths of one’s own forces. It seems foolish to me to try to compete with movies with our graphic displays, or novels with program code, instead of utilizing the one thing that they don’t have: processing power.
The second node is "human thinks". A game designer is tempted to write this one off; after all, that’s not my department, right? Well, no, it isn’t, but that doesn’t mean that we can ignore it. Exactly what is the human going to think about? What part of his brain are we going to stimulate? In the early years of the industry, we appealed exclusively to that least human of brain structures, the cerebellum. Later on, we decided that we would appeal to the analytic skills with puzzles and wargames. I suggest that there is a great deal more human mental activity that we have not yet touched in our games, and further that, if we can learn how to challenge the most personal aspects of human thought -- namely, intuition -- we will accomplish a great deal more with our games.
Sadly, the emphasis on technical expertise in our profession tends to mask the need for greater emphasis on human thought. Programmers who have memorized the syntax of every command in UNIX do not understand the fundamentals of linguistics. Designers who know all about the fine points of display techniques on CRTs do not know the difference between a rod and a cone. People who can talk all day about bus architectures blink when you mention the corpus callosum. I think it would behoove us, as mediators of the human/computer interaction, to balance our knowledge of both parties.
The third node is "human talks". This is the node that is, in so many games, the real bottleneck that plugs up the free flow of information through the circuit. Imagine a typical shoot-em-up computer game in progress: the computer is saying things like, ’big green spaceship flying through space, brilliant yellow fireballs arcing toward you, red and orange explosions, black space, stars...’ What does the human player get to say in response? ’Up’, ’down’, ’right’, ’left’, or ’fire’. How can you get any kind of rich interaction with so limited a set of inputs?
There are, of course, lots of excuses for not providing richer inputs to the human player. Perhaps the human is restricted to an input device such as a joystick that does not naturally say anything more than the aforementioned five verbs. There is also the serious problem that the human player must be able to learn any interface language that the game uses. It is much easier to understand the computer’s output than to master a new input system, especially when the former is visual and the latter is essentially tactile. These excuses, however, mean only that game designers must focus much more attention on the problem of facilitating the player’s input into the game. A properly designed game allows the player to say just as much back to the computer as it says to him. Ideally, the player would have exactly the same expressive language available to him as is available to the computer; this, however, is difficult to achieve in practice.
The fourth node in the interaction loop is "computer processes". This is the artificial intelligence used in the game, and it plays a vital role in closing the loop. The computer’s processing is the only thing that relates the player’s inputs to the computer’s outputs. That relationship must make sense, it must be interesting, and it must be believable. If I say to the computer-character Jezebel, "Jezebel, my love, I must do my duty on the field of honor," Jezebel had damn well better burst into tears. It will certainly require a hefty amount of processing for Jezebel to figure out the meaning of my statement and her reaction. Yet, without that processing, the loop is broken. Again, many games are weak in this area and suffer for it. The interaction loop is not the definitive or final means of assessing a computer game. It is a useful mental tool that can expose some of the flaws in a design. It suggests to me that a great many game designers would do better to concentrate less attention on the "computer talks" node and more attention on the other three.