|Microsoft DirectX 9.0|
Traditionally, applications have done their own mapping of events to particular buttons and axes. A car-racing game, for example, might assume that the x-axis on the user's joystick or mouse was the most suitable control for steering the car. The only way to accommodate new or unusual devices was to provide configuration options so that the user could specify some other axis, such as a rotational axis, to use for steering. Moreover, the application had no way of knowing which installed device was the best fit for the game, so the user typically had to choose a device from a menu or make sure only the preferred device was attached.
Using action mapping, you no longer need to make assumptions about the best use of devices and device objects. Instead, your application binds actions to virtual controls wherever possible. Rather than getting data from the x-axis and steering the car to the left or the right accordingly, the application might get data from a virtual control called DIAXIS_DRIVINGR_STEER. Microsoft?DirectInput?assigns the virtual control to a physical control—that is, a device object. It does so by taking into account the application genre, user preferences, information from the device manufacturer, and the user's configuration of the device.
Action mapping also simplifies the input loop by returning data for all devices in a form independent of the particular device. A single action can be mapped to more than one device, and the input loop can respond to the action the same way regardless of which device is being read.
The following topics contain more information about the steps required to implement action mapping.