Dear Ulf,
This set of problems is indeed very important to the history of computing, by my count. I recall a particularly powerful letter from Von Neumann
to Oswald Veblen, sent by Von Neumann during a 1943 visit to the UK [my emphasis added]:
“I think that I see clearly that the best course for me at present is to concentrate on Ordinance work, and the Gas Dynamical matters connected
therewith. I think I have learned here a good deal of experimental physics, particularly of the Gas Dynamical variety, and that I shall return a better and
impurer man. I have also developed an obscene interest in computational techniques. I am looking forward to discussing these matters with you. I really feel like proselytizing — even if I am going to tell you only things which
you have known much longer than I did.” (Quoted in full in Bill Aspray’s
John Von Neumann and the Origins of Modern Computing, 27. Original letter from 21 May 1943, Oswald Veblen Papers, Library of Congress).
That Von Neumann spent some time at the Nautical Almanac Office (L.J. Comrie’s old haunt) during this trip is, I think, significant.
I have recently enjoyed thinking about these matters along lines outlined by Robert Moir in his chapter “Feasible Computation: Methodological
Contributions from Computer Science” in Physical Perspectives on Computation, Computational Perspectives on Physics, ed. Michael E. Cuffaro and Samuel C. Fletcher (Cambridge: Cambridge University Press, 2018).
Perhaps these “feasible computations” are quite central to the history of modern computing, although I am not entirely convinced by Moir’s
claim that we can “trace the historical origins of the … approximate form of [feasible computation] to techniques developed by physicists to overcome the computational limitations of the mathematical formulation of theories and models of natural phenomenon.”
(see “Feasible Computation,” 174).
I’m afraid that I don’t possess the technical expertise to be of much more help in terms of 20th century “fluid dynamics,” but if your work
on feasible methods ever brings you back to the 17th century (as it did Herman Goldstine in his
History of Numerical Analysis), please drop me a line!
Best,
Patrick Graham
Graduate student at the University of Minnesota Program in the History of Science, Technology and Medicine