Every time I see that thing, I want to take a huge chunk of prime numbers and write code to visualize them in different ways as quickly as possible. How do they look if you put them in, say, a hexagonal grid instead of a square one? Do patterns emerge if you draw them on a Hilbert curve? A 3D Hilbert curve? 4D? Has anyone tried these? Am I crazy for wanting to? Why are the ducks watching me?
No! Do it! Fork my trivial program and put up your results.
The square is a pretty interesting place to start, though, since we already know there's a relationship between quadratic polynomials over the naturals and the distribution of primes. These polynomials describe lines that (eventually) become straight.
This is definitely amateur math hour for me. I was a math major a few years ago, but I did not study any number theory.
I'm leaning towards making something in Ruby, as I'm more familiar with it. Then I can throw a few dozen (hundred?) million pre-calculated primes at it, and look for macro-scale patterns, without having to do much calculation at all. Just plotting points.
Maybe it's time to dive deeper into Objective C and / or OpenGL. I've got a few ideas that would be fun to try in real-time, but most of the more "fun" programming languages have poor / slow graphics libraries, and something low level lets you handle millions of entries without breaking a sweat.
Regarding diving deeper into Obj-C/OpenGL, you could also consider prototyping your ideas quickly using something like Quartz Composer. I don't know any specifics on whether it'd deal properly with a large data-set, but the learning curve is not steep at all and it might get you going quickly.
I've seen QC, poked at it a couple times, but never in depth. Any recommendations for resources?
Usually I'm struck by how unbelievably awesome the UI is, and how scores of similar attempts have been incomprehensibly awful. Meanwhile, Apple comes out with QC, a fast, elegant, simple version that goes almost totally unnoticed. Weird.
