How I got acquainted with Phil Roe

I can say with certainty that, among my American colleagues in Computational Fluid Dynamics (CFD), I am the only one who knew Phil Roe before he became a CFD icon. Here’s the story.

Think summer, 1976. It was the year of the 5th International Conference on Numerical Methods in Fluid Mechanics, 6/28-7/3. My conference contribution, “Towards the Ultimate Conservative Difference Scheme IV,” had been rejected, as had been my contribution “Towards the Ultimate Conservative Difference Scheme III” to the previous conference of the series in 1974. That 4th conference was held in Boulder, CO, and since I had no paper to present, Leiden Observatory did not give me any travel money. Luckily, the 5th conference was held in my home country of the Netherlands, at the Technical University of Twente.

Several weeks before the meeting, the participants received a Book of Abstracts. The most intriguing one was from an unknown British author, who reported he had found a way of discretizing the characteristic flow equations compatible with the conservation form.

The conference went by without the British chap showing up. I was really disappointed and wrote to him at the provided author’s address, requesting a copy of his conference paper. I received no answer. After a month, I sent him a picture postcard of the windmill “De Valk” in Leiden, repeating the request, again without success. I concluded the author had found an error in his paper and was too embarrassed to reply. As for myself, I gave up submitting manuscripts to this international conference series, as I would never get through the selection process. The Dutch were allotted only two papers, and the local organizing committee would always select papers on low-speed problems, like flow around a wooden shoe.

Fast-forward four years to the 7th conference in the series at Stanford/NASA Ames, 6/23-6/27, 1980. My circumstances had changed dramatically. I had spent a sabbatical year at ICASE, NASA Langley, had built myself a network of colleagues in numerical math and CFD, and had obtained funding from Langley’s Theoretical Aerodynamics branch, then headed by Jerry South, for a second sabbatical year.

I had no paper to present, just came to listen. On the second day of the conference, at the end of the morning session, listening to a speaker with a British accent, I realized that his presentation, with the modest title “The use of the Riemann problem in finite difference schemes”, was getting increasingly interesting and ultimately exciting. After the talk, there were a few tame questions; I was the only one rushing to the podium to meet the speaker. “Hi, I am Bram van Leer. That was a most interesting presentation; may I have a copy of your paper?” Instead of introducing himself, the speaker said: “I must apologize to you for not having answered your letters twice!” He was the British chap of the 1976 abstract! He told me his name, Philip Roe, long escaped from my memory, never to be forgotten again.

He related what had happened in 1976. Because he had not yet presented his paper in the UK, he had to ask his employer, the Royal Aircraft Establishment (RAE), for permission to present it abroad. The RAE, suddenly realizing the potential of Phil’s work, misguidedly decided to classify it. Phil tried to reverse the designation, but, not a particularly organized person (as I later learned), he must have acted late, as the final decision on the paper was to be made the day before he would travel. It so happened that the person with authority was out of the office that day; thus, the default decision remained in place.

And so it remained for four years. Phil was not to make any mention of a paper that contained the basis of his discovery: fluctuation splitting. Meanwhile, he was anxiously watching my slow stream of JCP articles, two in 1977, one in 1979, expecting the next one to make his work obsolete. Because of that possibility, RAE finally allowed him to speak about it in public.

Having this cleared up, we left the lecture room, left the building, left the whole conference behind, found a shaded spot with garden furniture, and sat down. There we talked for seven hours straight, two able minds in tune. A meeting of historic consequence.