nawl

Conference, this week, at DIMACS. Take all of the mathematicians mentioned in, say, The Man Who Loved Only Numbers and throw them in a lecture hall with a handful of grad students. Well, only the living ones. Erdos might as well have been there, though, for all that he was incarnated.

Laszlo Lovasz on convergent sequences of graphs, for the second time. Throwing away all the details, imagine that you define some notion of distance between two graphs. Then you can have a Cauchy sequence of graphs. If the number of vertices goes to infinity, what does the graph converge to? Why, a function defined on the unit square. Isn't that neat? Now look at the problem from the other direction, and imagine that you are creating a random graph as follows: the connection matrix is a sort of discretization of a function defined on the unit square, where an edge between x and y is created with probability f(x,y). Of course this makes sense. Paper here though I haven't read it.

A very small and beautiful talk by Rados Radoicic on rainbow arithmetic progressions, where a rainbow arithmetic progression means each term in the progression is colored with a different color. A counterexample, this, I think, recursively generated by writing 1 2 * 1 2 *... and replacing each * with a term of that same sequence.

DIMACS is housed at Rutgers, which meant that I was waking up at 7 am to catch a train and then a bus to where I needed to be on the campus. Rutgers, of course, is the ultimate big state school, a university of something like 40,000 students sprawled out across a huge swath of New Brunswick. According to the class schedule, there are 20 minute breaks between classes to allow students to get from one part of the university to another on buses.

And staring out the window of the campus bus, I was somehow reminded of an eastern bloc country. It was something about the ill-advised 60s architecture, the blocky brick highrises, the wide streets along flat green countryside, the inexplicable 6-lane freeway separating one part of campus from another, the huge stadium complexes. But at the same time it's very American, the girls are broad-shouldered and dyed-blonde in flip-flops, each one outfitted with iPod nano and cell phone, chatting cheerily with friends about last night's drunken exploits.

Peter Shor came to speak at Princeton last night, and to have lunch with grad students today. (For those keeping score at home, he was the one who came up with an algorithm for factoring in polynomial time on a quantum computer.) I know nothing about quantum computation. I realized only last night when it was pointed out that the problems that are known to have faster quantum algorithms are those (like factoring) that are generally believed not to be NP-complete. Why is this exciting? Well, the idea that computational power is a physical property of our universe. Could you ask a string theorist to come up with a computational model? "Probably, though nobody understands string theory and the ones who do are too busy trying to understand the nature of the universe." I had to defend my discipline: "But we compute things precisely to understand the nature of the universe."

Well, apparently quantum information theory is going somewhere.

For amusement there's this. If only everything could be so amusing. And while I'm at it, I have to express appreciation for this poem, which I must confess I've had a fondness for ever since it (or something similar) was projected on the wall of a certain Berkeley apartment so many years ago.

I have been doing other happy things too, in all the time that I'm not spending sleeping. Buying plane tickets, planning things. Very complicated. In a few days I will perhaps post a schedule.

(The flowers, by the way, are all within a few minutes' walk of the grad college.)