Today was the first day of the 2002 meeting of the Stanford Computer Forum. This is a group of tech companies with Silicon Valley presence that recruit students and sponsor research at Stanford. At the annual meeting, scientists from all the Forum companies come to Stanford to interact with Stanford students and researchers and each other.
The software wasn't quite ready, though, so Costa and I spent until almost 3:00 polishing it and finding some nagging bugs. Hey, it's amazing how filenames with colons in them can screw up software that uses a colon as a field delimiter. But we fixed that and it was ready to go not a minute too soon, seeing as the demo was supposed to start at 3:00.
We showed it to one or two groups of people soon after 3:00, but the real surprise for me was when Monica walked along leading none other than Don Knuth himself, my greatest idol in computer science. I demonstrated it for him and we talked about its state and the future possibilities for maybe as long as ten minutes. He gave at least one suggestion of a feature that hadn't occurred to me before. This was the first time I met him and I was not disappointed. He was the height that a great man should be—at least two inches taller than me—a quality that I've often missed in other of my role models.
The demonstration period ended at 5:30, and I was faced with a decision. I could run off quickly home and go to taekwondo practice at 6:00—this is probably what I should have done—or I could stay for the Forum banquet. In the end I dallied too long talking and the decision was really made for me. I headed off to the banquet.
It turned out to be a good decision, whether voluntary or not. The highlight for me was another chat that I had with Don[*]. I asked him about the way he went about writing TAoCP (batch: he has a database of 20,000 articles collected over the years and categorized by section, and before he writes each section he reads all of them in its section very carefully) and about his experiences with offering rewards for bug-finding (very good, and it greatly speeds the rate at which errors are reported; about half of them are cashed), and we talked about binary tree and balanced trees for a while (he spent a long time implementing AVL trees in a filesystem circa 1965 and it sounded like he had problems debugging his deletion routine that were similar to mine).
After dinner, Larry Page from Google talked about his ideas about research and how it should be done. He was introduced by Terry Winograd, whom as it turns out was Larry's advisor at Stanford (from which Larry is "on leave"). I got a free Google baseball cap, one of many scattered around the banquet room.
I'm looking forward to the second day of the Forum meeting tomorrow, as well as the second seminars on Friday.
[*] I find it hard to call him by first name, but we're very informal in this department. Monica was surprised when I referred to him as "Dr. Knuth" this afternoon.