OXO

OXO is a computer implementation of noughts and crosses (tic-tac-toe) written for the Electronic Delay Storage Automatic Calculator? at the University of Cambridge around 1952 by the PhD student Sandy Douglas.⁷ It ranks among the first video games created in the early history of the medium, produced in an era when the games that existed were confined almost entirely to research laboratories run by universities, corporations, and governments rather than made for public consumption.⁴

The game let a human opponent play a full game of noughts and crosses against the machine; moves were entered using a telephone dial, one of the EDSAC’s input devices.⁷¹ The board and play were displayed on the machine’s monitors, exploiting a feature of the EDSAC’s serial memory whereby the contents of the store could be shown on cathode ray tube (CRT) monitors.⁷ Three of the EDSAC’s monitor tubes were visible in period photographs of the machine.¹ An image of the game served as the cover illustration for the tutorial guide to the EDSAC simulator, where it is described as an interactive computer game developed by a student programmer in 1952.¹

The EDSAC platform

OXO existed only on the EDSAC, described as the world’s first stored-program computer to operate a regular computing service.⁷ Designed and built at the Cambridge mathematical laboratory under the direction of M. V. Wilkes, with W. Renwick as co-designer, the EDSAC performed its first fully automatic calculation on May 6, 1949.⁷¹ The machine was a physically enormous device: it drew press coverage as a “mechanical brain,” was reported to contain roughly 3,500 valves, weighed about a ton, and occupied some 500 square feet across 120 racks assembled from early in 1946.¹

In the late 1940s the EDSAC — and “electronic brains” in general — captured the public imagination and were widely reported in the British press, from the Daily Mail to the Daily Telegraph and The Star.¹ Wilkes told the Daily Mail in October 1947 that “the brain will carry out mathematical research” and might even “solve economic and philosophic problems too complicated for the human mind,” while other reports speculated that a dozen such machines “would probably be sufficient for the whole country”.¹ It was against this backdrop of fascination with mechanical minds that a program setting a human against the computer in a game acquired its resonance.¹

The EDSAC’s processor occupied the bulk of the machine — some 3,500 electronic tubes in all — while its memory was built from mercury delay lines, or “tanks”.¹ The main memory was designed for 32 delay lines each storing 32 words of 18 bits, giving a total capacity equivalent to about two kilobytes.¹ Input and output were handled by a five-track paper-tape reader running at 50 characters per second and a Creed teleprinter running at 6⅔ characters per second.¹ A useful property of this early serial memory was that the contents of the store could be displayed on cathode ray tube monitors, three of which stood at the back of the machine.⁷ It was on this primitive hardware, with almost no facilities for troubleshooting, that Douglas’s game ran.⁷

Context and contemporaries

OXO belongs to a small cluster of games written for first-generation computers around 1950.⁴ Historians treat “first” claims about these programs with caution, since the games of the period were largely confined to research labs and many were likely lost without wider exposure.⁴ The earliest known computer game actually implemented is generally taken to be Bertie the Brain?, a custom-built tic-tac-toe machine devised by Josef Kates and unveiled at the Canadian National Exhibition in the summer of 1950 to promote his Additron vacuum tube; the thirteen-foot device used a lighted keypad and panel and was dismantled at the exhibition’s end.² Bertie could counter a player’s move nearly instantaneously and was virtually unbeatable at its highest difficulty, which Kates would lower for children and raise again for adults.²

A near-contemporary was Christopher Strachey’s draughts program, written for the Manchester and Ferranti Mark I? computers.⁶ Dated papers preserved in the Bodleian Library indicate it was mainly developed in June and July 1952, with a first version probably written before May 1951, and it very probably constitutes the first use of a graphical display in a computer program, showing the draughts board on the machine’s storage cathode ray tubes.⁶ Strachey demonstrated the program at the ACM conference in Toronto in September 1952.⁶ The draughts program implemented a heuristic look-ahead strategy, and Strachey has been credited with writing what may be the first heuristic program.⁶

These early games shared a common fate: because they stayed locked in research laboratories and were usually dismantled or discarded once they had served their purpose, they rarely spread beyond a few academics and exerted little to no direct influence on later games.⁴ Among video-game histories, the 1950s have largely been passed over, with detailed treatment confined to a few works such as Tristan Donovan’s Replay and Mark Wolf’s The Video Game Explosion.⁴

OXO itself survives chiefly through documentation and through the EDSAC Simulator, a faithful emulation of the original machine developed at the University of Warwick by Martin Campbell-Kelly, whose distribution bundles OXO among its demonstration programs.¹⁷ The simulator reproduces all the controls and displays of the original machine and includes a library of original programs, subroutines, debugging software, and documentation, so that users can experience what programming a first-generation computer was like.⁷ In it the game is opened from a folder of demonstration programs, run with the Start button, and played by entering moves on the emulated telephone dial.⁷¹ Later editions of the simulator have been associated with the EDSAC Replica Project at The National Museum of Computing at Bletchley Park.⁷