Nim is a very simple combinatorial game that was the focus of several early experiments related to computer and video games, because an automatic opponent can be so easily implemented. At the beginning of a round there is a certain number of objects—usually matches—from which players take turns to remove a certain number. Depending on the rules played, the object is either to be the player to remove the last object, or force the other player to do so (i.e., leaving only one object on the table for the other player’s turn). Commonly played setups include the objects being divided amongst 3 piles and players taking as many objects as they want from one single pile per turn; or all objects (usually 21) being on a single pile with players taking between 1 and 3 per turn.

Once understood, the game’s predictability means that it is easy to play a perfect strategy by always striving for “safe positions”. Given a specific initial setup and both players knowing this strategy, the matter of which player will win is simply a question of which player is allowed to take the first turn. While this simplicity led to Nim being the first game that humans were able to play against an automatic device, it also means that it loses its charm quite quickly. A player either knows the perfect strategy, in which case the games are entirely predictable, or they don’t, in which case they will always lose until they learn it.

It is the simple mathematical description of the game that has led Dr. Redheffer, as he states in his 1948 article in The American Mathematician Monthly (Vol. 55 No. 6, Jun/Jul 1948), to seek to implement it in an electronic circuit. The resulting device, using only electrical switches, was presented at various science fairs and open house exhibitions.

All but forgotten after the early 1950s, it seems that American toy designer and collector Mike Mozart was the one to rediscover the actual device in 2010, and determine its historical significance. While not identified beyond any doubt, the exact consistence between the device and Dr. Redheffer’s 1948 article, as well as the lack of almost any information about it before Mozart’s discovery, seem to suggest that it is in fact the original “machine for playing the game Nim” built at MIT in 1941-1942.

« Return to Part I — Prehistory

References

• Wikipedia: First video game
• Original article in The American Mathematician Monthly 55.6 (1948; JSTOR access required)

In his 1936 paper On Computable Numbers, with an Application to the Entscheidungsproblem, English mathematician Alan Mathison Turing introduced the concept of a simple, hypothetical “automatic machine”, later named a (universal) Turing machine. He proved such a machine to be capable of computing anything that is computable, using only a read/write head performing elementary operations (read, write, move) over an infinite memory, governed by strictly following a set of atomic rules. Since Turing showed that all computable algorithms can be reduced to such a set of rules for his conceptual machines, they serve as the basic theoretical model for a digital general-purpose computer—the ultimate hypothetical computing machine. Such machines are said to be Turing-complete, a label that applies to all modern computers.

Turing applied these ideas to show that there can be no general solution to the Entscheidungsproblem, which asks for an algorithm that can determine truth or falsity of any mathematical statement. He did so by showing that his conceptual machines are not able to solve the equivalent halting problem, which asks for an automatic machine to determine whether any other automatic machine, given its program (i.e. set of rules), will ever terminate. (The gist of it: in order to examine the machine and program to be tested, the testing machine would essentially need to emulate, or re-enact, all the steps it performs—so, if the process being examined never reaches a conclusion, neither will the process that examines it.)

These important theoretical results, and their very practical applications to the design of digital computers and programming languages, have earned Alan Turing the honourable title of “father of computer science”. Interestingly, few people recognised the importance of the work at the time of its publication. The Turing Award, the so-called “Nobel Prize of computing”, is named in his honour, and awarded annually to the most influential computer scientists.

In 1937, at the Massachusetts Institute of Technology (MIT), Claude Elwood Shannon completed his master’s thesis, of which a portion was published as a paper entitled A Symbolic Analysis of Relay and Switching Circuits in 1938. In it, Shannon proposed to consider relay switching networks as Boolean expressions, where open and closed relays represented Boolean values of 0 and 1, respectively. Boolean algebra could then be used to simplify existing arrangements, or find the most efficient implementation of new requirements. However, this is overshadowed by the implications of the reverse operation: modelling Boolean expressions as networks of relays, which could then be used to solve logic problems. This is the foundation of digital circuit design, and by extension all of electronic digital computing.

The same principle was already described two years prior, in a 1935 thesis by Russian electrical engineer Victor Shestakov. However, Shestakov’s work was not published until 1941, so the credit for creating a sound theoretical basis for digital circuit design went to Shannon instead. Experimental designs in digital circuitry had been realised earlier, but in a rather limited and ad hoc fasion. Shannon’s way of thinking paved the way for more complex logic circuits, translated to vacuum tubes and transistors, eventually leading to today’s incredibly complex integrated circuits (ICs).

Shannon continued to expand on his ideas, earning the title “father of information theory” by almost single-handedly creating the eponymous new field of science with the publication of another paper in 1948 (a separate article on the topic will follow).

Alan Turing and Claude Shannon knew each other, having met and exchanged ideas on numerous occasions.

« Return to Part I — Prehistory

References

• Wikipedia: Alan Turing, Turing machine, Turing Award
• Wikipedia: Claude Shannon, Shannon’s paper
• Portrait of Alan Turing: MacTutor History of Mathematics
• Portrait of Claude Shannon: MacTutor History of Mathematics

Early, pre-electronic arcade machines came in many different forms. People could use a penny scale to learn their weight, or receive a printed truism from a fortune teller machine. Strength testers encouraged friendly physical competition, while automatic musical instruments and early jukeboxes allowed to set the mood for a price. Some machines sold postcards and stamps, ball gum, hot peanuts, or cigarettes. Others combined several popular functions, such as scales that would also tell your fortune, or earn you your penny back should you have guessed your weight correctly. And let’s not forget slot machines or the early electro-mechanical pinball games either.

The imagination of manufacturers seemed limitless, and hundreds of new machines were released every year in North America and Western Europe. Some, such as the Mills Novelty Co. out of Chicago or the Caille Bros. in Detroit, had decades of great financial success, while countless others had to close shop after selling only a handful of their first design. So why mention 1934 in particular? Because of two machines that are documented to have been released this year, which more than those before them remind us of the arcade videogames that would start to appear almost 40 years later: a racing game and a flight simulator.

Based in Paris, France, manufacturer Autoreflex made their debut with a machine called Le parcours automobile, which quite possibly is the first racing game ever created. The machine’s cabinet housed a roll of paper, painted with a road and obstacles. A small model car was mounted on a rod that ran across on top of it. The car could be moved left and right by turning a big steering wheel mounted in front. After inserting a coin, the paper roll would start turning, and the player had to move the car to keep it on the road and avoid any obstacles. The same, simple principle was implemented by the early racing videogames.

It is not currently documented if or how the machine determined how well the player was driving, if there was any feedback or consequences when leaving the road or hitting obstacles. A few machines are said to still exist and be in the possession of collectors though. In any case, the groundwork for one of the earliest and most enduring videogame genres was laid. Autoreflex even released a sequel called Autoroute sportive three years later, which operated on the same principle and definitely recorded how many accidents a driver got themselves into.

Another one of the early, enduring genres is that of the flight simulator (although these games have, due to their complexity, always been more popular on home systems and home computers than in the arcades). Not much is known about Learn to Fly, a rather complex arcade machine released by an unknown manufacturer, also in 1934. Instead of a toy car, a small airplane was suspended in a glass box. A joystick, a throttle lever, and rudder pedals provide the realistic controls. When a coin was dropped, the glass box actually turned into a miniature wind tunnel, which remained active for a certain amount of time. The player controlled the plane according to instructions on the machine, with the plane said to have reacted realistically.

Apart from being the first flight simulator game designed for the public, Learn to Fly was quite possibly also the very first joystick-controlled game in general.

« Return to Part I — Prehistory

References

• Le parcours automobile (1934) at Arcade History and the International Arcade Museum.
• Autoroute sportive (1934) at Arcade History and the International Arcade Museum.
• Learn to Fly (1934) at Arcade History.
• Find more information on classic arcade amusement machines at Arcade History and the International Arcade Museum (previously KLOV – The Killer List Of Videogames).
• Picture: Library of Congress (Public domain)

Recording wires are made out of steel or stainless steel, and are only about as thick as a human hair. They are pulled past the recording head at a very high speed when compared to tape recorders, with the electrical signal of the sound supplied to the head and magnetising the wire as it passes. Playing back the wire without applying a signal to the head leads to the magnetic fields on the wire inducing the signal back into the head, the recorded sound can be reproduced. Despite the high speeds of recording and playback, the thinness of the wires allows for very high capacities. Wire recorders were thus sometimes able to record several hours of audio non-stop, in contrast to similarly-sized tape recorders.

The initially complex, heavy, and expensive magnetic tape recorders soon caught up to become a viable alternative for home use, and almost completely replaced wire recording during the 1950s. Magnetic tapes, first patented by Fritz Pfleumer in 1928, went on to become one of the main methods of data storage for home computers in the 1980s. But the principles first employed by Poulsen also lead to drum memories, floppy disks, and hard disk drives. Even in his 1898 patent, Poulson described that instead of a wire, one might use “a disk of magnetisable material over which the electromagnet may be conducted spirally,” or “a strip of some insulating material such as paper [covered] with a magnetisable metallic dust,” essentially predicting hard disks and magnetic strips as used on tickets or credit cards. Today, magnetic media continue to be the predominant forms of permanent computer data storage, to be rivalled only in certain applications by optical or semiconductor technologies.

Note: The Electrical World magazine issue of September 8, 1888, printed a description of magnetic recording by Oberlin Smith, who himself claimed to know about the principle from a visit to the Edison laboratories 10 years prior. However, it is not known whether any working model was ever built before Poulsen’s patented Telegraphone in 1898. In addition, Smith believed a continuous steel wire would not allow for precise magnetisation, and suggested regular wires furnished with steel dust.

« Return to Part I — Prehistory

References

• Wikipedia: Valdemar Poulsen, Wire recording, Fritz Pfleumer
• Magnetic recording history, University of San Diego (archived)
• 100 Years of Magnetic Recording, AMPS newsletter issue 27
• Picture: “Magnetone Magnetic Wire Recorder” by Radio Rover (CC 2.0 by-nc-nd)

While there are two fundamentally different methods of operation for a CRT, the physical principle remains the same. A cathode emits a constant stream of electrons, forming a ray, which is deflected and directed towards a screen coated with a fluorescent material. Wherever the electrons hit, the screen lights up.

In the first method of operation, rays are deflected to directly trace out the shapes that are to be displayed on screen. This is the way of CRT oscilloscopes, as well as vector displays, which have been used for a number of arcade videogames and less successful home game systems.

The other method has rays illuminating the whole screen surface by repeatedly tracing out a raster pattern, line by line. These screens are known as raster displays. The patterns to be drawn are controlled by modifying the intensity of the electron beam rather than its deflection—more electrons means a brighter dot on the screen. Raster displays also give meaning to the notion of screen resolution and refresh rate, or frame rate, commonly measured in frames per second and measuring how often this raster pattern is traced out each second. Higher frame rates reduce update delays, jerkiness of motion, and screen flicker. Colour displays have three beams separately amplified and targetting separate raster locations for each of red, green, and blue.

Up until the advent of flat-screen display technology about a hundred years later, CRT monitors were ubiquitous in television receivers and computer monitors, meaning that if computer and video games were played, they were most likely played on a CRT.

« Return to Part I — Prehistory

References

• Wikipedia: Cathode ray tube
• Picture by Interiot, Wikimedia Commons (CC 3.0-by-sa)

What both of these projects had in common was that they were not realised, primarily due to a lack of funding. The first person to successfully create a punch-card-based information storage and retrieval system was Herman Hollerith, who presented his Electric Tabulating System in 1884 and finished the project in 1886.

The motivation to commercialise his system was a government contract. Analysing data and preparing results for the decennial United States Census took, for its last run on data for 1880, eight years to complete, because all the work was still being done manually. Extrapolations suggested that, due to the growth in population, the 1890 census would not be completed in time before processing of the 1900 census would have to begin. Thus the U.S. government held out the prospect of a contract to the person or company who would be able to provide them with the fastest means to automate the process.

According to his own account, Herman Hollerith was inspired by observing railroad conductors, who used to encode a crude description of a passenger in the way they punched holes into their tickets, in order to prevent several people from sharing a ticket. The main element of Hollerith’s system was a tabulating machine with a card-feed mechanism which, fed with stacks of punched cards, could accumulate data and print out results automatically and accurately.

Hollerith won the contract, and the performance of his machine exceeded all expectations. The 1890 census was tabulated within only one year, with a total population count being released to the press after only six weeks. The success allowed Hollerith to found his own business, the Tabulating Machine Company, in 1896. Merging with three other corporations, it would become the Computing-Tabulating-Recording Corporation (CTR) in 1911, before being renamed to International Business Machines (IBM) in 1924. Piggybacking on their punched card system, IBM rose to be the predominant manufacturer of large-scale business computers for most of the 20th century. The concept of storing data on punched cards was the first application of computers to information processing instead of just mathematical computations, a step whose importance cannot be overemphasised.

For an interesting and touching read on computing with punched cards in the 1970s, I recommend to you Dale Fisk’s Programming With Punched Cards, written in 2005 and set predominantly in 1973.

« Return to Part I — Prehistory

References

• Wikipedia: Herman Hollerith
• Wikipedia: Punched card
• Wikipedia: 1890 United States Census
• Wikipedia: Semen Korsakov (includes animations)
• The Census of 1900 (National Geographic, January 1900)
• Dale Fisk – Programming With Punched Cards (2005)
• Picture “Hollerith punched card” (Railroad Gazette 1895): Wikimedia Commons (Public Domain)
• Picture of Herman Hollerith (Charles Milton, 1888): Library of Congress, adaptation: Wikimedia Commons (Public Domain)

How closely does it resemble the architecture of a modern computer? The purely mechanical and steam-powered device (no electrical parts involved) would have included modules for data input and output, a memory unit, as well as an algorithmic processing unit. Programs, data to work on, and store/load instructions would have been fed to the machine on punched cards through three separate readers. Card reading motion was reversible, and the machine supported libraries of subroutines. Apart from a printer, output could be sent to a curve plotter or a card puncher, as well as triggering a signalling bell.

While the algorithmic unit, called the mill, would have operated on registers storing 40-digit numbers, the memory, also called the store, would have provided storage space for 1’000 50-digit numbers, which is equivalent to about 20.7 kB. Apart from basic arithmetic operations, the Assembly-like microcode language for the mill supported loops, as well as conditional jumps based on the outcome of number comparisons, which would in fact have made Babbage’s engine Turing-complete.

While the general public at the time was not able to comprehend Babbage’s ideas, his scientist contemporaries mostly thought that he was a dreamer who had lost touch with the realm of the possible. One notable and famous exception was Augusta Ada Byron King, Countess of Lovelace, a young mathematician who said of herself to be “fascinated by the universality of his ideas.” Italian Federico Luigi, Conte Menabrea, wrote a report about Babbage presenting the concept of his machine in 1842. Augusta Ada King subsequently started translating his report from French to English. As King had already been a personal friend to Charles Babbage for many years, and followed the development of the Analytical Engine for the past 10 years, Babbage asked of her to add her own thoughts as notes to the translation—the notes ended up being three times as long as the original report.

These notes, which have long since become computer science history, show how deep King’s understanding of Babbage’s ideas actually went. Apart from explaining some of the mathematical background left out of the report for conciseness, her more interesting contribution were thoughts about the operation and application of such a machine, should it ever be realised. She realised that while previous, single-purpose calculation machines required more manual interaction of the user, an automatic general-purpose computer in turn posed an increased responsibility on its programmer to put care into the design of his input. Computer scientists today know this dilemma as computers allowing us to “make the same mistake many times, really fast.”

And even more astonishing is the fact that Augusta Ada King was probably the first person to consider such a device to not only be used for science and calculations, but also art and entertainment. While even Charles Babbage himself seemingly did not see his machine as being of use to anyone but mathematicians, King noted that the machine does not operate on numbers per se, but symbols, and that if any other kind of data were to be translated into symbols to which the machine’s operations were applicable, the machine could work on it as well. She adds that if, for example, the “fundamental relations of pitched sounds in the science of harmony and of musical composition were susceptible of such expression and adaptations, the engine might compose elaborate and scientific pieces of music of any degree of complexity or extent.” This extension of Babbage’s ideas to the arts and, in fact, general data and information processing, seems almost prophetic from today’s point of view.

A point of debate is whether the program included in her notes, an algorithm to compute the Bernoulli numbers using the Analytical Engine, was actually written or only debugged by her. Belief that the former was true has earned Augusta Ada King recognition as the first real computer programmer, and has led to the U.S. Department of Defense naming a programming language in her honour—ADA. In an autobiographic work however, Babbage states that he wrote the program and King merely pointed out a mistake that he had made, before adding it to her notes. On the other hand, Babbage is said to have been very miserly in giving credit to others. Whatever the case may be, the fact remains that Augusta Ada King was an exceptionally intelligent and insightful mind, an unknowing and largely unheeded herald of the computer age.

And so was Charles Babbage himself. Researching his original plans, historians and engineers came to the conclusion that Babbage would probably have been able to fully realise his Analytical Engine even with the manufacturing processes of his time, had he been able to overcome his problems in financing the project. Charles Babbage never gave up on the idea that one day, his vision would be realised: “As soon as an Analytical Engine exists, it will necessarily guide the future course of the science.”

« Back to Part I — Prehistory

References

• The History of Computing Project Timeline: 1830-1844
• Wikipedia: Analytical engine
• Wikipedia: Federico Luigi, Conte Menabrea
• Wikipedia: Ada Lovelace
• The History of Computing Project: Ada Byron King
• L.F. Menabrea – Sketch of the Analytical Engine Invented by Charles Babbage, with Notes Upon the Memoir by the Translator Ada Augusta, Countess of Lovelace
• Picture “Charles Babbage, obituary portrait”: Wikimedia Commons (Public Domain)
• Picture “Ada Augusta Byron King” by Margaret Carpenter: Wikimedia Commons (Public Domain)

Among those are punched cards as a method for storing programs for sequence-controlled machines. At the time, sequence control was commonly used for textile looms, where weaving patterns were defined using perforated paper rolls. Jean-Baptiste Falcon and Basile Bouchon improved upon this concept in 1728 by building a loom that was programmed using wooden punched cards, tied together with strings. This more robust and essentially modular way of storing programs caught on, was adapted to other devices and improved upon by other inventors, such as Joseph Marie Jacquard with his own loom head design in 1801. It was an especially big step forward compared to other established methods of storing control sequences, such as the tedious knotted ropes, or expensive to produce cog cylinders.

Punched cards proved to be a very successful concept. A couple of decades after Jacquard’s improvements, in the 1830s, Charles Babbage was planning to use them to program his unrealised Analytical Engine, and Herman Hollerith applied punched cards to the problem of storing data rather than programs towards the end of the 19th century. The cards were a primary way of data storage for computers far into the 20th century, to be replaced only by magnetic storage methods.

Another technology that was experimented with heavily, and fruitfully, over the course of the 18th century was the electrical telegraph. The tempting vision of sending messages over wires at the speed of light, a revolution in long-distance communication, saw many parallel strands of development. Early proposals would rely on one wire for each letter in the alphabet. It would be some decades into the 19th century until successful systems would be implemented on a usable scale, most notably the Morse telegraph by Samuel Morse and Alfred Vail in the U.S., along with the Morse code, which can be reduced to a binary code to be transmitted over a single wire.

This first form of electrical communication allowed instant transmission of messages over long distances and even between continents. The electrical telegraph is a direct predecessor to telephones and computer networks, the importance of which for the history of computers and computer games does not need to be spelled out.

As an amusing side note, 1769 saw the first presentation of a widely popular attraction built by a Hungarian inventor, Johann Wolfgang Ritter von Kempelen: The Turk, the first chess-playing automaton. Paying patrons were given the chance to play a game of chess against a humanoid machine that actually used its arms to move the pieces. Of course, the secret behind this was something less anachronistic than the first computer chess algorithm.

The ingeniously designed playing table actually housed a compartment for a human chess player, who was able to observe the contestant’s moves through magnets fitted under the chessboard. The operator in the compartment also controlled the automaton’s response by using an elaborate mechanical arm on an internal representation of the chessboard, which would translate into the figure moving its arm to carry out the moves on the outside. Details such as visible, non-functional machinery parts, a clockwork ticking sound while The Turk was “thinking”, as well as the operator winding up the machine with a key, supported the illusion. Proof that long before computers, people liked the idea of playing games against non-human, automatic opponents.

« Back to Part I — Prehistory

References

• The History of Computing Project Timeline: 1672-1770
• Wikipedia: Punched card
• Wikipedia: Jacquard loom
• Wikipedia: Electrical telegraph
• Wikipedia: The Turk
• Wikipedia: Wolfgang von Kempelen
• Picture “Jacquard loom”: Wikimedia Commons (Picture by Mahlum, Public Domain)
• Picture “Collins Overland Telegraph Cabin on Lake” by John Clayton White: Wikimedia Commons (Public Domain)
• Picture “The Turk” by Joseph Racknitz: Wikimedia Commons (Public Domain)

One of the most impressive applications of these concepts was probably Heron’s automatic theatre: a 10-minute stage play automatically controlled and performed using such a programmable machine, including the triggering of special effects such as the sound of thunder.

Sequence control is a central principle to the concept of electronic games. It is what allows a machine to present to players an opposition, in the form of changing obstacles to overcome or seemingly intelligent opponents, providing the challenge that leads to fun and a sense of accomplishment. Similarly, it allows automatic enforcement of rules in puzzle-style games, or games for several players. On a very basic level, sequence control is what separates electronic games from traditional games.

Coincidentally, Heron is also believed to be the creator of the first automatic, coin-controlled vending machine, to be placed in temples. An inserted coin fell onto a pan, which would be lowered due to its weight. The tilting pan would open a lever, causing holy water to be dispensed to the customer. The pan continued to tilt until the coin fell off, thus closing the lever and stopping the flow of holy water after a certain amount had been dispensed. The combination of such vending machines with automated games would lead to coin-operated amusement devices and, almost two millennia later, Arcade video games, which offered patrons a few minutes of entertainment in exchange for their hard-earned coins.

« Go back to Part I — Prehistory

References:

• Wikipedia
• Picture: MacTutor History of Mathematics