Alistaire Cooke died yesterday at age 95. I was introduced to Cooke through his Letter to America, the long-running radio series that ended only a couple weeks ago. I enjoyed his radio program, and I have enjoyed his writing. I only hope I survive to such energetic old age.

I spoke at the local matrix computations seminar today. I stayed up too late last night working on the slides, and my presentation was unpolished, but the talk went acceptably well. Whether or not my audience learned anything new from my talk, I learned something truly remarkable from it. In preparing my talk, I tried to describe in pictures why a certain phenomena occurs. When I described the picture during the talk, I realized that I had not carried my idea as far as it would go. After lunch with colleagues, I walked home, thinking as I went. I paused at the French Hotel Cafe and wrote down my idea, then wrote about a different problem for a while. I finished my walk; put my room in order and checked e-mail; had dinner, tea, and an interesting conversation about Hamilton's principle with Dave; talked to Winnie; and then returned to my insight from earlier in the day.

I wrote the programs I needed in about twenty minutes. Neither one was complicated. I tested them on a small problem, and they worked beautifully. They were fast, too, and solved most of my problem in a fraction of the time taken by approaches I'd tried until now. So I gleefully set up a larger problem and set my programs to work -- and the network connection died.

That's okay. The network will come back tomorrow. It was still a good day.