The FPU(T) Problem

The study credited with including the first computational experiment, which is also marked as the birth of my field of nonlinear science (although there is a very long prehistory because of the work of Poincaré, etc.) is of the so-called Fermi-Pasta-Ulam (FPU) problem, which consists of a set of masses connected by nonlinear springs. The scientists were trying to study the idea of thermalization/ergodicity from statistical physics (a problem that goes back to Debye) in which an initial excitation of one mode would ultimately lead to an equipartition of energy in the system. Of course, things turned out to be much more complicated than they were expecting, and that's why over 50 years after their technical report appeared in 1955, a lot of research is still being conducted on this and similar systems. (By the way, my work on chains of granular materials uses an FPU-like model.)

One of the interesting things I noticed several years ago when going through the original paper---in addition to the article's Brooksian title and the fact that this article has the only abstract I've ever seen that includes the term "untimely death"---was the footnote, "We thank Miss Mary Tsingou for efficient coding of the problems for running the computations on the Los Alamos MANIAC machine." I always wondered who Mary Tsingou actually was because I could never find any other papers by her. So I was naturally very excited when this article came out in Physics Today in January and answered my question. (As you can now see, this is yet another delayed blog entry.) It turns out that she changed her name in her scientific work after getting married, and indeed continued to work on related scientific problems. Unfortunately, her fundamental contributions to this problem (which included doing the heavy lifting in the original paper!) have been largely forgotten. I'm not sure how things work with technical reports---note that the original paper wasn't actually published in a refereed journal and I still don't know the story of why Tsingou wasn't listed as a coauthor---but if this were one of my papers, she would have been first author. The Physics Today article proposes changing the name of the problem, though we'll see how well it can fight 50 years of habit and scientific literature.

By the way, another very interesting thing in the article is the preparatory diagram for the computational experiments and the story behind how those experiments found the amazing phenomena that have captured the imagination of so many people (including me). One could now do this in something like 15 lines of vectorized code.

In conclusion, don't forget the following: It's a MANIAC, MANIAC (on the floor)!

P.S. If you can't access the version of the article in Physics Today, it has been posted on the arXiv.