I just saw the news that Benoît Mandelbrot died a few days ago (14th Oct)
http://www.nytimes.com/2010/10/17/us/17mandelbrot.html
I think it is fair to say that it was my first introduction to fractals, and writing my first program in BASIC on my ZX Spectrum in around 1985 (age 12), that opened my eyes to the beauty of complexity in the natural world.
It was then that I suddenly appreciated that the deterministic nature of physics as described by maths means even if a system seems random, or too complicated to understand. It is still driven by the same simple rules. From that day I new my path would be in physics.
So if something as complex as the Mandelbrot set can be described by the simple mapping of z->z^2+c, then the rest of the Universe may not be beyond our grasp.
That was 25 years ago…
Now, I find myself working on trying to describe the nature of turbulence in plasmas, where the self-similar (ie. fractal) appearance of eddys over a range of scales can tell you about how energy is distributed and dissipated.