Donald B. Gillies


Donald Bruce Gillies was a Canadian computer scientist and mathematician who worked in the fields of computer design, game theory, and minicomputer programming environments.

Early life and education

Donald B. Gillies was born in Toronto, Ontario, Canada, to John Zachariah Gillies and Anne Isabelle Douglas MacQueen.
He attended the University of Toronto Schools, a laboratory school originally affiliated with the University. Students at this Ontario school skipped a year ahead and so he finished his 13th-grade studies at the age of 18. Gillies attended the University of Toronto, intending to major in Languages. He started his first semester taking seven different language courses. In his second semester he quickly switched back to majoring in Mathematics which was his love while in high school.
During his time as an undergraduate, he spent a great deal of time at the U-Toronto Computation Center. In the Putnam exam competition of 1950, Gillies was stunned at not being selected by the faculty to compete with the U-Toronto team. To avenge himself, Gillies placed in the top 10 in North America, following his University of Toronto classmates John P. Mayberry and Richard J. Semple who were top 5 Putnam Fellows. Toronto would likely have won the competition in 1950 had Gillies been on the faculty-designated team.
For graduate school, Gillies applied to the University of Illinois University of Illinois at Urbana-Champaign, which was "a very busy place building lots of computers". While he was there, he began working on the ORDVAC/Illiac I project. After one year of graduate school, Gillies transferred to Princeton University to work with John von Neumann, at the urging of and also to be with John P. Mayberry, who was also studying under John von Neumann. Gillies and Mayberry were both arch-rivals and best friends, and after Mayberry beat Gillies in the Putnam exam, each competed to finish his PhD degree first.
During his graduate studies, and after working with von Neumann, Gillies became a fan of the book "One-upmanship" by Stephen Potter. John von Neumann was also a fan of this work, and was extremely successful at impressing others with his intelligence. An apocryphal math problem asks about a bumble bee flying back and forth between two approaching trains, and how far did it fly before colliding? When von Neumann gave the correct answer, the questioner asked if he used a standard time/rate-of-travel trick, and he replied, 'no, I summed the infinite series in my head' to impress the questioner. This method of impressing and astonishing others appealed to both Gillies and von Neumann.
During his time at Princeton his interest area was computer design first and mathematics second. He continued to work with U-Illinois researchers and participated the check-out of the ORDVAC Computer at Aberdeen Proving Ground in Maryland, in the summer of 1951.
At one point during his graduate studies, von Neumann found out that Gillies had been spending time working on an assembler. Von Neumann became enraged and told Gillies to stop work immediately because computers would never be used to perform such menial tasks.
After only two years of study at Princeton, Gillies completed his PhD before Mayberry, at age 25, in 1953, which was published in Contributions to the theory of games, vol.2 — in which he characterized the core which is the set of stable solutions in a non-zero-sum game.

Early career

Gillies then went to England for two years to work for the NRDC and worked with an early Ferranti Pegasus computer there. This was a time when the American, British, and Canadian, governments were conscripting young men for service in the Korean War. In later years, after he returned to the U.S., he was again drafted, but successfully appealed against the order.
While at the NRDC, Gillies and Christopher Strachey filed several American, British, and Canadian patents.
The patent on Order Control laid out the details of how to implement a base register for program relocation in computers - before it had been done. He considered these patents as kind of a joke, and assigned the rights of the patents to either NRDC or IBM, without taking fees for this service. This kept the ideas from being patented by others which would have hindered progress in the computer industry.
When Gillies returned to the US in 1956, he received a 1-A draft status which persisted until he was age 36. Upon returning to the US, Gillies married Alice E. Dunkle and began a job as a professor at the University of Illinois at Urbana-Champaign.
In early October 1957 the Soviet Military launched Sputnik I, and caused a widespread panic across the United States. Just hours later the UIUC Astronomy Department rigged an ad-hoc interferometer to measure signals from the satellite. The astronomers approached Dr. Gillies and Dr Jim Snyder to program the ILLIAC I computer to calculate the satellite orbit from this data. The programming and calculation was completed in under two days. The very rapid publication of the ephemeris in the journal Nature—just a month after satellite launch—helped to dispel some of the fear created by the Sputnik launch by the Soviet Union. It also lent credence to the idea that the Sputnik launch was part of an organized effort to dominate space.
Starting in 1958, Gillies designed the 3-stage pipeline control of the ILLIAC II supercomputer at the University of Illinois. The control circuitry consisted of advanced control, delayed control, and interplay. This work was in the public domain, and competed with the Stretch computer system design from IBM that is often credited with inventing pipelining. This work was presented in a 1962 Michigan conference on computer design, "On the design of a very high speed computer"
As the main designer of the pipelined control circuitry for ILLIAC II, Gillies developed the algorithms for the month-long checkout and acceptance testing of the new computer. To draw attention to this new computer design in the field of mathematics, he wrote an implementation of the Lucas–Lehmer primality test and found three new Mersenne primes, and published them in a paper, "Three new Mersenne primes and a statistical theory."
The new Mersenne primes were reported in the Guinness Book of World Records, and the largest one was immortalized on all mail sent from the Post Office at the Math department of the University of Illinois. In the same paper, Gillies made a conjecture about the distribution of prime divisors of Mersenne numbers.

Later career

In the late 1960s, Gillies became concerned that students were not getting direct access to computers any more. He lobbied UIUC to adopt the 1968 WATFOR one-pass FORTRAN compiler / runtime system from the University of Waterloo in Ontario. This was a fast-turnaround IDE for batch-based mainframe computers. At the time it was common practice to submit a job and pick up the results the next day. The WATFOR compiler could compile, link, and run a short program in the compiler's memory space in a few seconds. This compiler allowed the university to offer undergraduate programming courses not only to computer scientists but also to business majors and to other non-specialists. Gillies and his family traveled to Waterloo to pick up a magtape with this compiler, on one of his visits to see his family, in the early 1970s.
In 1969, Gillies received a preprint of Wirth's "Pascal User Manual and Report" and launched a project to build the first Pascal compiler written in North America. Ian Stocks was one of the graduate students who worked on this fast-turnaround in-memory 2-pass compiler, and the compiler was completed in the early 1970s. This work was part of the "PDP-11 Playpen" project which focused on getting graduate students direct access to low-cost computer hardware, such as the PDP-11/23, where the Pascal compiler ran.
Two years later at the urging of his new graduate student, Greg Chesson, Gillies became in 1974 the first licensee for the UNIX operating system from Bell Labs.
Chesson went on to be the third person to edit the Unix kernel and was the eighth hire at Silicon Graphics Inc.

Personal life

Gillies met his future wife, Alice E. Dunkle, while at Princeton and began dating her, but after several months, their relationship fizzled. Miss Dunkle, knowing of the rivalry between Mayberry and Gillies, intentionally flirted with Mayberry at a dance, and Mayberry subsequently approached Gillies to ask if he was still dating her. This tactic, used only once, led to their eventual marriage.

Death and legacy

Gillies died unexpectedly at age 46 on July 17, 1975, of a rare viral myocarditis. Digital Equipment Corporation and many of his friends, colleagues, and family contributed money for the Donald B. Gillies Memorial Lectureship In Computer Science, at the University of Illinois at Urbana Champaign. This annual lectureship continues to this day.
In 1994, the Nobel Memorial Prize in Economic Sciences was awarded to John Forbes Nash. In the Nash Seminar, Gillies was mentioned as a pioneer in the field of game theory. Nash proved the existence of stable solutions for non-zero-sum games; Gillies and Shapley extended this work by characterizing the core which is the set of stable solutions that cannot be improved by a coalition.
In 2006 the Donald B. Gillies Chair Professorship was established in the department of Computer Science at the University of Illinois. A generous donation from Lawrence White, a former student, established this chair. The first professor to hold this chair is Lui Sha, a well-known authority on real-time and embedded systems.
In 2011, the UIUC Department of Computer Science awarded a Memorial Achievement Award to Gillies, and family members accepted the award on his behalf at the Urbana-Champaign campus.
In 2018, the Donald B. Gillies Chair Professorship endowment had grown so large that Vikram Adve was invested as the second chair professor at UIUC under this designation. Adve led the project that developed the LLVM compiler suite which has been adopted industrywide.

Students