
Bertil Gustafsson Oral History (pdf) Interviewer: Philip J. Davis Gustafsson discusses his career in numerical analysis and applied mathematics in both Sweden and the United States. Gustafsson's interest in mathematics began early. As six yearold boy living in rural Sweden, he discovered his talent in mathematics thanks to a perceptive schoolteacher. While lucky enough to have several good teachers, Gustafsson believes that being a good mathematician requires talent as well. He believes his success in numerical analysis is mostly attributable to hard work and a little luck. Gustafsson received his Ph.D. from Uppsala. When HeinzOtto Kreiss arrived at Uppsala in the early 1960s, Gustafsson was in his second year, had already developed a strong preference for applied mathematics, and so quickly took to Kreiss, his work with numerical analysis, and the new field of computing. His dissertation was on difference methods for hyperbolic equations. Gustafsson has continued to pursue topics in differential equations for most of his career, and much of his early work consisted of applying his theoretical developments to practical problems. These included weather prediction, both in Sweden and at the National Center for Atmospheric Research in Boulder, Colorado and in a variety of problems in various fields during a oneyear postdoc as an itinerant consultant searching for useful applications of differential equations in Swedish industry. Currently, Gustafsson is working with David Gottlieb on a variety of problems, which until recently included work in computer tomography. Gustafsson, who often works with engineers, notes a certain tension between this group and mathematicians when the two work together, in which neither fully understands the challenges of the other's job. Part of the problem in working with industry is the question of proprietary versus open knowledge, and Gustafsson has encountered this issue often with his work at the Parallel and Scientific Computing Institute in Uppsala. Gustafsson believes that teaching while researching helps provide feedback for one's research, and so can be quite valuable. He suggests that incoming mathematics Ph.D. Students need a solid background in general mathematics, and notes that his program in Sweden requires students to have some experience in working with computers and to have taken a few courses in numerical analysis. Key words: difference equations, hyperbolic equations, partial differential equations, weather prediction, Swedish industry, computed tomography Funding Agency: NASA, NOAA, NSF Time frame: 1960's, 1970's, 1980's People: HeinzOtto Kreiss, David Gottlieb, James Ortega, Joseph Oliger, Gene Golub Location: Uppsala University, NCAR, Parallel and Scientific Computing Institute Uppsala, ICASE, California Institute of Technology, Stanford Citation: Bertil Gustafsson, Oral history interview by Philip Davis, 21 August , 2003, Brown University, Providence, RI. Society for Industrial and Applied Mathematics, Philadelphia, PA Statement of Use Policy: Copyright © by the Computer History Museum. Use of this the material for research purposes is allowed. Any such use should cite the SIAM History of Numerical Analysis and Scientific Computing Project (http://history.siam.org). Use of the oral history materials for commercial purposes requires the written permission of the Computer History Museum. Contact the Computer History Museum, 1401 N Shoreline Boulevard, Mountain View, CA 940431311 USA for permissions. 
Copyright ©, Society
for Industrial and Applied Mathematics This project is funded by a grant from the United States Department of Energy. 