I consider control to be the last of the constraint categories based off the physical characteristics of play. Also, it is one of the few “universal” constraints I might talk about, as it is applied to any kind of games. I would even go as far as stating that most games are determined by some form of constrained controls.
If space determines how many game elements the game can hold and how they can be arranged, it is by control that a game action is carried out, changing the state of one or several game elements.
Control possibilities affect the way we interact with a given game system. Changing the associated control scheme (a subset of input possibilities) of a given game modifies it directly. The nature of this modification can be objective (the rule system is modified) and/or subjective (our perception of the game is modified).
Every game has an incompressible associated control scheme, where each possible recognized input offered by the scheme has an associated game object or state. In mathematical terms, the relation between controls and game objects is injective, i.e. for every element of the control domain, there is at most one associated element in the game domain, while there can be elements in the game domain with no corresponding element in the control domain.
Okay, so I might have gotten a little bit too enthusiastic with the math there. Basically, it’s a fancy way to say that a game’s control scheme allows you to interact with a limited number of game objects at a given moment. Games never give full control of game objects to a single player, because then it stops being a game to become a toy.
Video games are a very good laboratory to study constrained control, as inputs are noise-free, whereas in the physical world, passive non-game objects can interfere with the accomplishment of a game action. There are two main archetypes of input methods: continuous (mouses and joysticks) and discrete (buttons).
(Actually, “continuous” control methods are not really continuous, they are discrete input methods with many different possible states. However for simplicity’s sake, we will admit that from a human point of view, these devices appear to be continuous.)
But since real life is continuous, why even bother with those limited, boring old buttons, then? There’s a very good reason to that. The main benefit of discrete inputs is that they are very simple to understand. Our brains are very skilled at handling binary logic, not so much fuzzy logic. Imagine you want to fire a gun in a shooting game. Compare these two trails of thought: “if I press this, I will fire” and “the more I push on this thing, the more likely I am to fire”. Which one seems simpler?
Furthermore, early computer games have shown that continuity and simulated continuity are barely indistinguishable. All arcade joysticks, like the one shown above, had in fact only eight possible directions. To move around in the original Doom, you only needed the four arrow keys. In every modern FPS game on the PC, as complex and extense as they are, you get a very fine control of avatar movement by using only four buttons (the staple WASD keys) and one analog input. In Halo 3, you control your avatar with two analog inputs.
Again, the path I’m walking here has already been tread. It has been extensively demonstrated that most games can be controlled by using only one discrete input. I could point you to a fine Gamasutra article on “One Button Games” from some years ago, or the recent GameCarreerGuide Game Design Challenge on one button shooters.
My point here is that control is an effective and primordial constraint indissociable from the concept of “game”. The very moment we accept to limit our own actions to the actions allowed by a system of rules to achieve a non-trivial goal, we are playing a game.
And yes, I’ve been reading Huizinga.