An input device capable of returning precise positional information, usually in two dimensions.
Design
Shortly after the appearance of
videogames, the most common form of
controller took inspiration from
aeronautics. Originally, the
joysticks in
aircraft were mechanically linked directly to the
control surfaces such as the
ailerons,
elevators, and
rudder. As such, the control may be considered as
analog, incorporating more than a
digital on/off determination, but a measurable quantity of 'how far'. However, most early
arcade game designs required motion in only limited directions, and as such the joystick could be very simple in design, consisting of
microswitches in either 2 or 4
cardinal directions depending on whether the joystick was designed for movement in one or two dimensions. Such joysticks are essentially digital input devices, and, depending on casing, can be restricted to signal only one switch at a time, or allow two, for
diagonal movement. This was sufficient for most arcade game designs, but was a long way from the aircraft flight controls that inspired them. Analog designs were however inherently expensive, not so much in construction as in maintenance, so most initial videogames were designed with digital controllers in mind.
Original IBM PC compatible joysticks were typically of the analog design, since the input provided by such a controller could easily be downgraded to simulate a digital controller by setting a threshold at which the analog input could be treated as a digital 1, while of course analog games would be able to use the input directly. The hardware design of such a device is reasonably simple. Along each axis, the position of the joystick mechanically moves the slider on a potentiometer. This variable resistance is connected to a capacitor to form an R-C circuit. By first charging the capacitor, then measuring the time taken for the capacitor to discharge through the resistance below a particular level (commonly determined with a zener diode), the discharge time can be used as an indirect measurement of the resistance and hence the position of the joystick along that axis. The granularity and accuracy of such a measurement depends on the construction of the potentiometer and the linearity of its scale, hence good joysticks were typically expensive, while all joysticks of this construction (good or bad) tended to be easily destroyed. Other notorious weaknesses of the design were the mechanical spring couplings that provided the user with simple force feedback and returned a stick to its neutral position, and the dampers that restricted the stick motion to within technical specification.
Limitations
The most economical way to interface the PC to this device turned out to be the bane of the early PC game developers. The simplest digital interface was to draw power from a PC port, and use bitfields to determine whether the capacitor was charging, or whether it had completed its discharge. In this way the analog signal can be measured by digital operations, by sampling the discharge bit. Early products that ran under an operating system such as DOS had little choice but to set the first bit, and busy wait for the second, consuming valuable processor time that could have been dedicated to the game. As the operating system evolved to support preemptive multitasking, such an operation could be performed as a background task by the device driver. As hardware costs were reduced, some designs moved the sampling process out to an analog to digital converter on the interface card itself.
The software sampling of the analog value, indirectly supplied by the capacitor discharge time, led to a few technical issues on the early DOS games. Essentially, the discharge time was measured by counting iterations of a busy loop, and so was dependent on the clock speed of the processor, while more generally the lack of a standard for joystick hardware (not to mention the variations in supposedly-identical components due to an attempt to keep hardware costs down) meant a PC game typically had to calibrate the joystick early in the process, by taking measurements of the loop counter at the extremes of travel (such as top-left and bottom-right). At the same time, console systems such as the Super Nintendo Entertainment System and Sega Genesis were gaining in popularity, and controllers such as the Gravis Gamepad appeared for the PC, providing the same D-pad digital directional control. It appeared that the analog joystick would be useful for flight simulators and other similar games requiring precision motion, but most entertainment products did not need them. Economy then dictated a considerable amount of game design from then on.
It took the next generation of console development to reverse the trend. As systems grew in power to the levels of the Nintendo 64 and Sony PlayStation, 3D games became technically possible, and with them the need for more precise control. The analog joystick once again became the norm, although no longer in its original proportions. The primary directional control on such systems is a tiny analog joystick operated by the thumb, and most modern console controllers include two of these, one for each hand. While technology and manufacturing processes have changed to allow the mass production of these components, the underlying principle is the same. In many cases the D-pad is still present, but serves as a poor substitute for the Control Stick for most videogames, and indeed has found alternative uses in most modern game design.