George Szekeres 1911-2005 Aug 28

It is not often realized how "computational" George was, in addition to his work in analysis and combinatorics. The early history of George and Esther has been well told by Ren Potts [1]. There is a nice story about how on appointing George as lecturer, Adelaide University thought it would be a good idea if he got a PhD. Apparently the regulations would allow him to be his own supervisor and examiner, but not to be enrolled as he was insufficiently qualified for this. Before this could be sorted out, he had become professor at UNSW. George was already writing Fortran (or FORTRAN as it then was) programs around 1960, (when the language was just a few years old) for his multidimensional continued fraction algorithm. In the mid-90s, I worked with him on Feigenbaum's functional equation. We both wrote programs to solve several cases of this equation, and I was very impressed by this 80 year-old who knew more about how to actually get computers to do real mathematics than many of my younger colleagues. I thus became his last co-author. After this I corresponded with him on simultaneous Diophantine approximation algorithms, and his letters were always characterized by carefully worked-out illustrative examples. One of his last papers [4] was computational in a different way - he studied the growth rate of solutions of functional equations by ingenious Fourier methods.

biography

References

• [1] R B Potts, Mathematics at the University of Adelaide, 1944-58 Aust Math Soc Gazette 12, 25-30 (1985)
• [2] 90th birthday
• [3] move to Adelaide
• [4] Abel's Equation and Regular Growth: Variations on a Theme by Abel
• [5] Bibliography
• [6] death