By Michael Steuerwalt
Seymour Victor Parter, 1927-2022. SIAM photo.
Seymour Victor Parter, professor emeritus of mathematics and computer science at the University of Wisconsin–Madison and a past president of SIAM, died peacefully at home in Madison, Wis., on October 21, 2022. He was 95 years old.
Seymour’s career did not follow the usual track, in that he worked full time while earning his degrees. His father died early, so he and his mother ran the family restaurant; he worked there for 30 or more hours each week until it was sold in 1951. During that period, Seymour graduated from the Illinois Institute of Technology with a bachelor’s degree in mathematics and a minor in physics (in 1949), followed by a master’s degree in mathematics (in 1951). Karl Menger directed his master’s thesis in topology, which focused on generalized metric spaces.
Upon Menger’s recommendation, Seymour applied for a position at Los Alamos Scientific Laboratory (now Los Alamos National Laboratory), excited by the prospect of working on physics and math problems in the mountains. He was hired by Preston Hammer in March 1951. As a staff member, Seymour programmed numerical algorithms for machines like the IBM CPC and IBM 602A. Doing so introduced him to the type of serious scientific computing work that he would continue at Los Alamos, then later at the National Bureau of Standards (now the National Institute of Standards and Technology) and New York University (NYU). Through these efforts, Seymour befriended accomplished mathematicians and physicists such as Hans Bethe, Mark Kac, Peter Lax, Robert Richtmyer, Marshall Rosenbluth, Garrett Birkhoff, Stanislaw Ulam, John von Neumann, Edward Teller, Kurt Otto Friedrichs, and Louis Nirenberg.
In 1951, SEAC (Standards Eastern Automatic Computer)—a new computer that was designed and built by the National Bureau of Standards—became available to Los Alamos scientists. Seymour received a SEAC manual in October of that year and traveled to Washington, D.C., to run calculations on this more powerful machine. In May 1952, he returned to Los Alamos to work under Richtmyer in the Theoretical Division, where he programmed and ran important defense codes on a variety of machines. Staff member Lester Barnhoff taught Seymour how to be a human do loop for card-programmed computers when the computer memory couldn’t hold all program instructions simultaneously, or when the computer lacked a loop register. Seymour would read the computer program into the card reader until finding a colored card, which revealed how many times he had to run the next set of cards through the reader; to carefully count the number of times, he pushed pennies across a table.
In May 1953, Seymour was sent to NYU to work on the UNIVAC (Universal Automatic Computer), which had been recently installed by the U.S. Atomic Energy Commission. When Richtmyer told him that his stay would continue through the spring of 1954, Seymour enrolled in the graduate program at NYU’s Department of Mathematics (now the Courant Institute of Mathematical Sciences). He performed well, and Los Alamos agreed to assign him to UNIVAC projects each fall and spring so he could continue his classwork at NYU. Seymour was always grateful to Carson Mark, head of Los Alamos’ Theoretical Division at the time, for allowing him to complete his Ph.D. while maintaining employment.
Seymour wrote his graduate thesis, “On Mappings of Multiply Connected Domains by Solutions of Partial Differential Equations,” under the direction of Lipman Bers [1]. Although he finished his thesis in 1956, Seymour did not receive his degree until January 1958, as he had not yet completed the language requirement. Because the tests only occurred once a year and had to be passed six months before the degree was granted, Seymour had time for extracurricular activities while meeting this final obligation. He spent the 1957-1958 academic year at the Massachusetts Institute of Technology in what was essentially a postdoctoral appointment. Seymour considered this year to be his most important; he learned to ski and met his wife Ruth while horseback riding near Boulder, Colo. They married in October 1957.
Meanwhile, Seymour’s mathematical interests shifted from pure to applied and computational mathematics. The growing national enthusiasm for science, aerospace, and computation increased the demand in industry and academia for individuals with both mathematical and computational skills. Seymour became an assistant professor in mathematics and associate director of the Research Computing Center at Indiana University from 1958-1960, then moved to Cornell University from 1960-1962 and Stanford University from 1962-1963. In the fall of 1963, he accepted an associate professorship at UW–Madison with a joint appointment in the Department of Mathematics and Department of Numerical Analysis (now the Department of Computer Sciences). He remained there for 36 years but maintained his relationship with Los Alamos, partly because the Laboratory continued to generate interesting research problems and partly because he and Ruth had friends in Los Alamos and Santa Fe. He also oversaw the Ph.D. theses of 15 graduate students, at least five of whom worked at or visited the Laboratory.
Seymour’s research was broad, influential, and imbued with an awareness of associated computational issues and practical consequences. It involved the modeling, analysis, and computational solution of various physical problems, as well as the design and analysis of computational methods. Seymour was interested in complex analysis, singular perturbations, sparse matrix theory and its connection to graphs, numerical methods for partial differential equations (PDEs), iterative methods, and preconditioners for numerical solutions of matrix equations. Though he did not consider himself a specialist, Seymour made significant advances in multiple areas. For instance, he wrote several papers that linked graph theory to sparse elimination — including an early paper that was published in SIAM Review in 1961 [2]. Researchers still used the paper’s key lemma to design and analyze sparse elimination methods three decades later.
Seymour also conducted significant analysis on Toeplitz forms and the eigenvectors and singular values of Toeplitz matrices, which pertain to error estimation and preconditioning. His work in singular perturbations and bifurcation theory expanded to the study of chemical reactors and fluid flows between rotating disks. He authored several influential series of papers about block iterative methods for finite difference and finite element matrices, the convergence of multigrid schemes, and the norm and spectral equivalence of elliptic PDEs.
Seymour was active in the broader mathematical community and sat on numerous panels and committees, including an advisory panel to the National Academy of Sciences. He was also a prominent, engaged member of SIAM. He served on the Joint Projects Committee of the American Mathematical Society, Mathematical Association of America, and SIAM from 1978-1980; the SIAM Council from 1978-1981; and as SIAM President from 1981-1982. Seymour was also the managing editor of the SIAM Journal on Numerical Analysis from 1977-1980 and served on its editorial board from 1981-1997. In addition, he was on the editorial board for the SIAM Journal on Applied Mathematics from 1976-1979.
Beyond SIAM, Seymour became chairman of the Conference Board of Mathematical Sciences from 1983-1985. Furthermore, he edited books on numerical methods for PDEs, large-scale scientific computation, and (with Richard Meyer) singular perturbations and asymptotics. He was elected as a SIAM Fellow in 2009 and a fellow of the American Association for the Advancement of Science in 1992.
Seymour officially retired from UW–Madison in 1996, though he continued post-retirement service for three years and remained involved in probability seminars and topics courses for even longer. He and Ruth loved to travel and visited Germany, Australia, and New Zealand. Seymour kept attending conferences, giving lectures at universities, and conferring with other researchers about exciting problems. He also maintained two longstanding habits during retirement: enthusiastically participating in reading clubs and running with a group of friends along the lakefront near UW–Madison’s campus (slow enough for conversation but steady enough to reach an interesting destination, a practice that lasted until the runners were well into their 80s and 90s).
Seymour’s family—his wife Ruth, sons Paul and David, daughter-in-law Susan, and grandchildren Alicia, Ezra, and Danielle—as well as his intersecting circles of friends, colleagues, and students all agree that he was attentive, kind, caring, competent, and accomplished. He will be sorely missed.
References
[1] Parter, S.V. (1960). On mappings of multiply connected domains by solutions of partial differential equations. Commun. Pure Appl. Math., 13(2), 167-182.
[2] Parter, S. (1961). The use of linear graphs in Gauss elimination. SIAM Rev., 3(2), 119-130.
Michael Steuerwalt holds a B.E.E. and M.S. from the University of Dayton and a Ph.D. from the University of Wisconsin–Madison. He worked as a student and staff member at Los Alamos National Laboratory and was program director of the computational mathematics and applied mathematics programs in the National Science Foundation’s Division of Mathematical Sciences.