Ikeda Toshio was one of the main figures in the development of Japan's computer industry. He was also a keen amateur go player. He brought a computer engineer's logical analysis to the game itself, and to the question of its rules.

Ikeda was born in 1923 into an artistic family. His father and older sister were both painters. Ikeda's own interests were directed more toward music and mathematics, especially geometry. He taught himself higher mathematics in his early teens, won prizes in competitions, and devised original proofs of the Pythagorean theorem.

In 1946 Ikeda graduated from the Tokyo Institute of Technology and went to work for Fujitsu, Ltd. At that time Fujitsu's main business was manufacturing telephone equipment for the Ministry of Communications. Fujitsu's telephone sets did not always work perfectly, and at one point an order went out from General Headquarters (Japan was under American occupation) that Fujitsu was to cease production. Ikeda turned his talents toward finding the cause of the problem. This led to his first published paper, a thorough mathematical analysis of the rotary motion of a telephone dial.

Although the cease-production order was rescinded, this was a time of serious inflation in Japan and GHQ instituted stiff deflationary measures. One measure curbed expansion of the telephone network, and Fujitsu had to lay off a third of its work force. Seeking means of corporate survival, in 1952 Fujitsu's management decided to look into the new field of computers. In the course of experiments for his telephone-dial analysis Ikeda had built an electronic pulse counter, so it was natural for him to join the computer project.

Fujitsu's first computer, a relay-operated machine targeted at the Tokyo Stock Exchange, was designed and built in six months by a small team of engineers. Ikeda's performance on the team was sometimes erratic. He worked intensely on the design but forgot all about producing his assigned schematic diagrams. He was caught up in the theory of arithmetic and logic circuits, filling notebooks with this idea and that. During the test phase, when the team worked around the clock for two weeks, Ikeda kept it awake through cold nights by singing Schubert's "Winterreise" at the top of his lungs in German.

The completed computer had a balky temperament and caused its creators much anxiety. It worked when shown to the stock exchange officials, then broke down as soon as they left. The stock exchange rejected it in favor of a UNIVAC punched card system. Ikeda, however, had become irresistably fascinated by computers. A few months later he started designing the FACOM 100, a larger relay computer suitable for scientific computations.

Ikeda now proved to be a first-rate computer designer, but his working hours grew increasingly irregular. He had low blood pressure and was not at his best in the morning, so he would often toil through the night to the point of exhaustion, sleep during the day, then wake up in the evening and start in again. This was not the normal Fujitsu schedule, so Ikeda frequently worked at home. Fortunately, he had superiors who recognized his ability and tolerated his behavior.

Although only one FACOM 100 was ever built, it was followed by more advanced relay computers and peripherals and Ikeda continued to display his talents. An engineer was having trouble designing a linkage system to move the type bars of a line printer. Ikeda analyzed the problem, found a mathematical solution, and wrote a program for the FACOM 100 that quickly calculated the answer. This was Fujitsu's first experience with computer-aided design.

In the late 1950s Fujitsu began making vacuum-tube and parametron computers, followed in the 1960s by transistorized models. In his 1962 New Year address, Fujitsu's president declared that Fujitsu would stake its corporate destiny on computers. In 1968 the first FACOM 230-60 was installed at the University of Kyoto. This was a revolutionary machine with all integrated-circuit logic that supported multitasking and multiprocessor configurations. A best-seller, in 1970 it pushed Fujitsu into first place in market share among Japan's domestic computer manufacturers.

During this time Ikeda was promoted to positions of increasing responsibility, first in computer hardware then in software development. He is remembered not only for his engineering genius but also for his inspiring leadership qualities. Some idea of Ikeda's contributions can be gained from the list of awards he received:

in 1958, a Kanto Regional Invention Award for an "interconnection method for assuring the safety of electrical computer circuits;"

in 1958, a Mainichi Industrial Technology Award for a "relay-type electrical computer;"

in 1966, a National Invention Award for a "method of address modification in the address modification unit of an electrical computer;"

in 1966, an Ohm Award for "development of the FACOM 230-10 very-small-scale electronic computer;"

in 1970, a Merit Award from the Director General of the Science and Technology Agency for "development of the FACOM 230-60 large-scale electronic computer system;"

in 1970, an Imperial Invention Prize for an "interrupt system using an interrupt counter."

Ikeda's reputation as an engineer was matched by his reputation as an eccentric. He was the type of person who might be late to work because he had decided to listen to a Beethoven recording; who entertained dinner parties with juggling and balancing acts; who wrote about computers programmed to fall in love with each other and marry; and who, when scheduled to attend an important executive meeting, was once found at home instead, seated half-clad on a window ledge, carving a toy out of bamboo.

In 1968 a Fujitsu official arranged for Ikeda to meet Gene Amdahl. Ikeda had great confidence and pride in his engineering skills, but in the architect of System/360 he recognized an equal. After Amdahl set up his own company in 1970 the two men were able to associate more freely and soon became, in Amdahl's words, "fast friends." The relationship developed into more than just friendship. Ikeda had long wanted Fujitsu to establish an engineering center in the United States. In 1971 this dream materialized on the grounds of Amdahl Corporation in Sunnyvale, California. Next year Ikeda was able to persuade Fujitsu to invest financially in the new American company.

In 1973 Amdahl Corporation experienced growing pains, and Ikeda strove to tide them over. He shuttled by plane between Tokyo, where he convinced skeptical Fujitsu executives that the partnership with Amdahl was worth continuing; San Francisco, where he advised Amdahl on matters ranging from company management to hardware debugging and fought off suspicions that Fujitsu was attempting a take-over; and Chicago and New York where he lined up further support for the fledgling corporation.

International in his outlook, during this time Ikeda worked on joint projects in Spain and Canada as well. Then in November 1974 he suddenly collapsed while waiting to meet Canadian representatives at Tokyo's Haneda Airport. He died four days later of cerebral hemorrhage, without regaining consciousness. He was fifty-one years old.

Though he did not live to see the results, the efforts Ikeda made in the last year of his life turned out successfully. In December 1974 the first Amdahl 470V/6 crossed the Pacific, designed and developed in California, manufactured at Fujitsu plants in Kawasaki and Nagano, destined for installation at NASA. Other top-level mainframe users likewise began replacing their CPUs with the 470V/6, and Amdahl started marketing Fujitsu's new M Series of processors in the United States. This was what Ikeda had planned.

From childhood on, Ikeda enjoyed both intellectual and physical games. In high school he was a star center who led his basketball team to a hundred consecutive victories. This was also when he started playing go; he would stay up late at night after basketball practice to tackle the school Meijin. When Ikeda entered Fujitsu he played at the four-kyu level, but he was already keen enough to spend half his first month's salary on a magnetic go set. (That left him without enough money to eat meals; to fend off starvation he had to sell most of his books to a second-hand book shop.)

When back problems put an end to his basketball career, Ikeda's passion for go redoubled. He clipped the go columns from newspapers and amassed a huge collection of professional games, many of which he knew by heart. From 1950 to 1956 he and two other Fujitsu engineers took weekly lessons from a 2-dan professional. Always one to seek the strongest opponents, Ikeda later founded Seihokai, a group that met weekly with Go Seigen and Rin Kaiho. Awhile before his death Ikeda managed to win one game apiece from these two at a three-stone handicap.

When the Fujitsu group began planning its exhibit for the 1970 Exposition in Osaka, Ikeda was among the planners. The theme was "Computopia," a futuristic vision of the interplay between computer and man. Some of the interplay was to take place on the go board. Ikeda began looking for a set of rules logical enough to satisfy the requirements of computer go and, more importantly, the future international go-playing community.

Ikeda attacked the rules problem with the same intensity he had shown in attacking the rotary motion of the telephone dial and the design of computers. He filled notebooks with handwritten rule analyses, lists of principles, diagrams, comparison charts, and drafts of new proposals. He published his conclusions in a series of eleven articles that appeared between September 1968 and November 1969 in Igo Shincho.

In the course of these articles Ikeda revised his rules proposals several times. The final version, given in Part I of this book, consists of overlapping rulesets that share the same basic definitions and rules of play, but differ in how they end and score the game. Ikeda proposed three sets of Chinese-style area rules and two sets of Japanese-style territory rules. For completeness I have added a third set of territory rules.

The articles themselves form Part II. They are presented as they appeared in Igo Shincho with only minor editing, such as the insertion of a few parenthetical remarks and correction of mistakes in the diagrams. Some of Ikeda's comments about the Japanese rules are now out of date; updates are given in footnotes.

Ikeda's efforts at rule reform were largely ignored during his lifetime, but he would probably be glad to know that in 1978 the New Zealand Go Society switched to rules similar to his area rules II (differing mainly in allowing self-capture), and in 1991 the American Go Association formally adopted rules embracing both area rules II and territory rules III.

Three people have been most helpful in the preparation of this book. Okamoto Akira of Fujitsu Business Systems Ltd., who was closely associated with Ikeda as a computer engineer and a go player, provided the foregoing information about Ikeda's career. Fujisawa Yoshihiro of Fujitsu's Advertising Division made the publication arrangements. Okasaki Masahiro introduced me to Ikeda's work and provided copies of his articles and notes, as well as much other information and advice. I am grateful to all three, and to Miyamoto Naoki (9 dan) of the Kansai Kiin who gave permission to publish the material from Igo Shincho.

Japanese names are given in the Japanese order, family name first.

James Davies
December 1991