Colleagues Remember Germund Dahlquist
May 1, 2005
John Butcher, University of Auckland: �When I first met Germund Dahlquist 35 years ago, it was like visiting an old friend. His early work on consistency, stability and convergence was well known to me, and I fondly remember my enjoyment as I worked through the beautiful proof of the first Dahlquist barrier. When I met him in 1970, I started to appreciate that he was more than a brilliant mathematician and computational scientist: He was a kind and sensitive man and a loyal friend. To me he became a mentor and an inspiration. Over the years I tried to keep up with what he was doing as best I could. Especially notable was his introduction of G-stability and the later proof that it was equivalent to A-stability. Everything Germund published was a separate gem, exhibiting deep mathematical insight and, at the same time, a clear understanding of sound computational practice. He was a pioneer who remained a central figure throughout his career; he will be sadly missed.�
Peter Lax, Courant Institute of Mathematical Sciences, New York University: �I am a great admirer of Germund. His influence has been great and all to the good. I am also charmed by his sunny personality undarkened by the long Scandinavian winters.�
Wayne Enright, University of Toronto: �I have always viewed Dahlquist as being one of the world�s leading numerical analysts. In my discussions with him and the presentations I have seen him give, I was very impressed with his wide knowledge of numerical analysis across all sub-disciplines. In particular, he seems to have analysed and explored generic �general-purpose� approaches that are applicable across a variety of application areas and involve the development and analysis of an �enabling technology� rather than an application-specific advance or breakthrough. His 1963 BIT paper is an outstanding example where he defines and suggests a remedy for the difficulty of �stiffness� which had been recognized in different application areas, but could be (and was) addressed in an application-independent way.�
Arieh Iserles, Cambridge University, and Syvert P. N�rsett, Norwegian University of Science and Technology, Trondheim: �We take familiar things for granted. In particular, it is obvious to us that numerical practice is underpinned by solid, honest-to-god mathematical theory and this informs much of our professional life. This paradigm, which transcends any single theorem or result, we owe mainly to three individuals: Germund Dahlquist, Peter Lax, and Jim Wilkinson. In the early fifties they demonstrated that numerical algorithms do not just �happen�: They can be understood and must be justified by rigourous mathematical analysis. If, as numerical analysts, we can see so far today, it is because we are standing on the shoulders of these giants and their generation.
�In our personal contact we have found Germund to be always encouraging, inspiring, and far-sighted. In 1980 he was virtually the only numerical analyst to display enthusiasm toward an early precursor of Lie-group methods. After a talk by one of us in Stockholm in 1997, Germund told us: �I have several times tried to work in that direction, but was not able to get to the right results.� Knowing his mathematical prowess, we can only deduce that either he did not try hard enough or, as always, he was being modest, nice and encouraging.�
C. William Gear, Princeton University: �It is rare to meet a person who not only clearly enjoyed mathematics so much, but also enjoyed applying it to important problems of scientists and engineers at large, and had so much fun in doing it. I feel honored to have known Germund, have benefited enormously from his mathematical advances and deep insights, and will miss him as a friend.�
