“Meet the Computer Scientist You Should Thank for Your Smartphone’s Weather App”
by Sarah Witman
Smithsonian, June 16, 2017
Pitch: The ‘hidden figures’ who changed weather forecasting
I’m a freelance science writer based in Madison, Wisconsin (we met briefly at the last NASW meeting in San Antonio). Over the past few years, I’ve written about cybersecurity in the food industry for Quartz, HIV in Malawi for the Pacific Standard, spaceflight’s impact on the human body for Popular Science, and more.
I’m working on a story that I’d like to run by you. It would be a profile of Klara von Neumann and the other “hidden figures” who changed weather forecasting as we know it.
Generally speaking, modern meteorologists forecast the weather using a technique called numerical weather prediction: They input current conditions into computer models, which estimate what the weather will do in the future. This technique can be traced back to a single paper, published in 1950. A team of scientists — top minds in meteorology and computing — used one of the first computers, ENIAC, developed during World War II, to produce the first numerical weather prediction. Or, as the Weather Bureau put it, “These men had made the first successful barotropic forecast on a computer.”
Except it wasn’t just men. Klara von Neumann, who was married to one of the paper’s co-authors, acclaimed mathematician John von Neumann, was instrumental to the project. During World War II, Klara worked in Princeton’s Office of Population Research, calculating population trends. So when John started working on ENIAC after the war, she was well-equipped to become “one of the first ‘coders,’ a new occupation which is quite widespread today.” When the team convened in early 1950, it was Klara who “initiated them into the ways of the ENIAC and its peripheral card-processing machines” and “checked the final program.”
Yet, Klara is not mentioned in almost any papers or books about the history of weather forecasting. In the 1950 paper, she is simply listed in the acknowledgements, when the same work today would certainly earn her co-authorship. She is literally a footnote in history. The “staff of the computing laboratory of the Ballistic Research Laboratories” (which, at that time, was mostly women) also received an acknowledgement, though not by name, “for help in coding the problem for the ENIAC and for running the computations.” And Margaret Smagorinsky, Norma Gilbarg, and Ellen-Kristine Eliassen were not thanked for their “hundreds of hours of hand-computing” in trial runs before the ENIAC was up and running.
I would like this piece to focus on Klara’s contributions to this project, as described above, in the context of the many forgotten contributions of women in the early days of computing, which have only recently started to become well-known. I’ve interviewed Dr. John Knox, who teaches his students at the University of Georgia about Klara’s role in meteorology history, and Peter Smagorinsky, whose deceased parents both worked on the 1950 project.
Please let me know if this interests you. Looking forward to hearing your thoughts!
[The editor responded, asking for a few more details; my response is below.]
In regard to your question, I think there are two main things that set Klara, and this story, apart:
1. Not only was this the first computerized weather forecast, but it was the first time scientists had ever succeeded in using a computer to conduct a physics experiment. After the Manhattan Project, John von Neumann (possibly the most famous mathematician of that era – he’s thought to be the inspiration for Stanley Kubrick’s Dr. Strangelove) became interested in applying the ENIAC computer to do real-world good. This idea was looked down upon by his peers at first, who saw it as inferior to doing pure math, but the success of the experiment changed many of their attitudes. This paradigm shift might not have been possible without Klara’s contribution (described in more detail below).
2. While many women worked on the ENIAC, including on this experiment (Margaret Smagorinsky, the first female statistician for the Weather Bureau who was also married to meteorologist and co-author Joseph Smagorinsky, and two other women did all the coding for trial runs before the experiment, and “the staff of the Computing Laboratory of the Ballistic Research Laboratories,” pictured here in figure 4, co-wrote code and ran the computations for the full experiment), Klara got involved at the ground level, and had much more of a leadership role.
Being married to John von Neumann, she had the opportunity to work on ENIAC before it was even built. In the book Turing’s Cathedral, Klara recalls, “Long before [ENIAC] was finished, I became Johnny’s experimental rabbit… It was lots and lots of fun. I learned how to translate algebraic equations into numerical forms, which in turn then have to be put into machine language in the order in which the machine has to calculate it, either in sequence or going round and round, until it has finished with one part of the problem, and then go on some definite which-a-way, whatever seems to be right for it to do next… The machine would have to be told the whole story, given all the instructions of what it was expected to do at once, and then be permitted to be on its own until it ran out of instructions.” She says she found coding to be a “very amusing and rather intricate jigsaw puzzle.”
When the co-authors of the weather-forecasting experiment (Jule Charney, Ragnar Fjørtoft, George Platzman, Joseph Smagorinsky, and John Freeman) arrived on-site in March 1950, Klara “helped code their problem and initiated them into the ways of the ENIAC and its peripheral card-processing machines.” She also “check[ed] the final program.” According to Platzman’s letters to his wife, recounted in the book A Vast Machine, “Von Neumann himself rarely appeared, but called in frequently by telephone. Working around the clock for 33 days, the team at Aberdeen [including Klara] carried out two 12-hour and four 24-hour retrospective forecasts.”
To put the magnitude of Klara’s role in perspective, meteorology professor Dr. John Knox explained, “You had to do a lot of programming on ENIAC to get anything to work. Programming back then — the late 1940s — was not sitting down and writing code. It was about punch cards… There was no such thing as a screen back in those days. You had to actually punch out little holes that told the computer each little instruction. The number of cards for this particular problem was about 100,000. So guess who did that? Klara… When you have 100,000 cards, you have to make sure you don’t lose any of them… If one of them falls out of order, the whole program is screwed up… She was in charge of all of that.”
Let me know if you’d like to know more – thanks!