The first displays
The sciences that sit behind computer graphics are electrical engineering, electronics, and television, as those fields stood in the first half of the twentieth century. Screens had shown crafted images earlier than that. The Lumière brothers used mattes for special effects in films from 1895. Those displays were limited, and they were not interactive. Nothing you did in the room changed the picture, except walking out.
The first cathode-ray tube, the Braun tube, was invented in 1897. From it came the oscilloscope and the military control panel, which are the more direct ancestors of the field. They were the first two-dimensional electronic displays that answered to a program or to a person's input. Even so, computer graphics was barely a discipline until the 1950s. It emerged after the war from university and laboratory work on more advanced computers, and from military work on radar, aviation, and rocketry. Those projects produced more information than a row of dials could show. New displays were built to cope, and the discipline came with them.
Whirlwind, a light pen, a game of tennis
Early projects such as Whirlwind and SAGE made the cathode-ray tube a workable display and a way to interact, and they introduced the light pen as an input device. Douglas T. Ross, on the Whirlwind SAGE system, wrote a small program that captured the movement of his finger and displayed the vector of it, his traced name, on a scope. It is a modest demo. It is also the whole idea in one gesture: a hand moves, the machine records the path, the screen shows the path.
In 1958 William Higinbotham built Tennis for Two at Brookhaven National Laboratory, to entertain visitors. It ran on an oscilloscope and simulated a tennis match. It is one of the first interactive games with a picture a person could recognise and play. In 1959, still at MIT, Ross was working on turning mathematical statements into three-dimensional tool paths for a machine. In the course of that he put a Disney cartoon character on a display scope. The serious job and the picture were already sharing a machine.
Hewlett-Packard went public in 1957, having incorporated the decade before, and its founders' ties to Stanford were strong. That relationship is part of how the southern San Francisco Bay became the place later called Silicon Valley. Computer graphics grew up alongside the hardware business, not off to one side of it. A picture needed a display, a display needed circuits, and the circuits were a product.
Sketchpad
In 1959 the TX-2 was built at MIT's Lincoln Laboratory, with a set of new ways for a person and a machine to meet. A light pen could draw on the computer. Ivan Sutherland's Sketchpad, built on that machine around 1962 and demonstrated in 1963, is the program people still point at. With the pen you could draw simple shapes, save them, and call them back.
The pen had a small photoelectric cell in its tip. The cell emitted a pulse whenever the screen's electron gun fired directly at it. Time the pulse against the gun's current position, and you know where the pen is. The computer can then draw a cursor there. It is a simple trick, and it solved a problem that had looked awkward: the screen is a moving dot of light, not a surface you can touch, so "where is the pen" has to be asked in time rather than in place.
Sketchpad also changed what a drawing was. If you want a square, you do not have to nurse four lines into meeting. You say you want a box, and you say where and how big. The software constructs it. Better still, Sutherland's program modelled the objects themselves, rather than a finished picture of them. Given a model of a car, you could change the size of the tyres without affecting the rest of the car, or stretch the body without deforming the tyres. A lot of later drawing software is still cashing that cheque. Constraints, and a model rather than a finished mark, are the two ideas. Three-dimensional graphics starts from the second of them. Timothy Johnson's Sketchpad III, the same year, showed a perspective view beside top, front, and side views, as interactive wire-frame models. Hidden edges stayed visible. The picture was three-dimensional in its data and still a tangle of lines on a flat tube.
A name, a game, a company
In 1961 Steve Russell, another MIT student, wrote Spacewar! for the DEC PDP-1. Copies spread to other PDP-1 owners. DEC itself ended up with one, and engineers used the game as a diagnostic on every new machine before it shipped. Sales staff learned to switch it on for customers and call it the world's first video game. Higinbotham's tennis had beaten it by almost three years, but tennis had stayed inside a laboratory. Spacewar travelled with the computer.
Around 1961 and 1962, at the University of Cambridge, Elizabeth Waldram wrote code to display radio-astronomy maps on a cathode-ray tube. The picture was not a game and not a cartoon. It was a scientific instrument's output, drawn because a table of numbers was the wrong shape for the question. In 1963 E. E. Zajac at Bell Telephone Laboratories made a film, on an IBM 7090, showing how the attitude of a satellite could be changed as it orbited. Ken Knowlton, Frank Sinden, Ruth A. Weiss, and Michael Noll were working in the same laboratory. Sinden's film Force, Mass and Motion illustrated Newton's laws. At Lawrence Radiation Laboratory, Nelson Max made films of a viscous fluid and of shock waves. Boeing made a film of an aircraft vibrating. The early showreel of the field is a showreel of people explaining a physical system to someone who was not going to read the equations.
In the early 1960s Pierre Bézier at Renault used Paul de Casteljau's curves, now called Bézier curves, to model car bodies. Curves are harder to draw well than polygons. They are also what a body panel actually is. A great deal of later modelling rests on that shop-floor mathematics.
Companies arrived quickly. TRW, Lockheed-Georgia, General Electric, and Sperry Rand were among those active by the middle of the 1960s. IBM answered with the 2250 graphics terminal, the first commercially available graphics computer. In 1966 Ralph Baer, at Sanders Associates, devised a home game that Magnavox later sold as the Odyssey. It was simple, and the electronics were cheap. It let a player move points of light around a television. It was the first consumer product of the field.
David Evans had been director of engineering at Bendix's computer division from 1953 to 1962, then a visiting professor at Berkeley, interested in how computers and people meet. In 1966 the University of Utah hired him to form a computer science program, and graphics became his main subject. The department turned into the main research centre of the field through the 1970s.
In 1966 Sutherland built the first computer-controlled head-mounted display. It showed two wire-frame images, one per eye, so the viewer saw the scene in stereo. The hardware that held the display and the tracker was heavy enough to be called the Sword of Damocles, for the obvious reason. A head-mounted display aimed at virtual reality had already been patented in 1960. Sutherland's machine is the one that was driven by a computer and drawn as graphics. He went from MIT to ARPA, then to Harvard, and in 1967 Evans brought him to Utah.
In 1968 Evans and Sutherland founded the first company devoted to computer graphics hardware. Sutherland had wanted it in Cambridge, Massachusetts. Salt Lake City won, because the research group was there. The same year Arthur Appel, at IBM in Yorktown Heights, described the first ray-casting algorithm, the start of a family of methods that follow light from a source, to a surface, and into the camera. In 1969 the Association for Computing Machinery started SIGGRAPH, a special interest group for the field. The first annual conference was in 1973. It is still where a large part of the published work is argued over.
Two other names belong to these years because the work is specific. Jack Bresenham, in 1962, devised the line algorithm that still carries his name, and later the midpoint circle algorithm. He spent 27 years at IBM and then taught at Winthrop. Charles Csuri made what is counted as the first computer art in 1964. A. Michael Noll had been making digital art since 1962, using random processes on purpose, and in 1965 he, Frieder Nake, and Georg Nees were the first to exhibit computer art in public. That April the Howard Wise Gallery showed Noll's work alongside random-dot patterns by Bela Julesz. The field's first decade already contained a game, a weapons display, a car body, a satellite, and a gallery wall. It did not yet contain a picture most households would have recognised as coming from a computer.