There was an odd story in NA Digest a couple days ago, John Francis of QR found. When I saw that someone was found, I assumed he had lost as in lost at sea, like Jim Gray. But that wasn’t the case.

John Francis developed the QR algorithm, an algorithm for finding the eigenvalues and eigenvectors of a matrix. Some experts regard the QR algorithm as one of the 10 most important numerical algorithms of the 20th century. He developed the algorithm in 1959 but then left the numerical analysis community three years later. The NA Digest article doesn’t say whether Francis became a recluse or simply moved on to a job outside mathematics. No one in numerical analysis knew anything about him until a couple folks tracked him down recently. He is doing well. He remembers his earlier work clearly but was unaware of the impact it had had.

Related post: Simple legacy (how people often underestimate the importance of their most useful work)

The introduction to “The QR algorithm: 50 years later its genesis by John Francis and Vera Kublanovskaya and subsequent developments” by Golub and Uhlig [1] talks about their quest to find John Francis. It also contains a short bio.

Unfortunately, I don’t think the article is freely available.

I attended the 23rd Biennial Conference on Numerical Analysis, in Glasgow where John Francis gave a fascinating talk. It was amazing how well he remembered everything.

http://imajna.oxfordjournals.org/content/29/3/467.abstract