Generally it seems like you're just trying to throw as much "random stuff" into the pile as possible - but it's not really contributing anything.

Indeed. This reminds me of Donald Knuth's Algorithm K, the "super-random" number generator. He develops an complex 13-stage algorithm that is "intended to be so complicated that a person reading a listing of it without explanatory comments wouldn't know what the program was doing".

Sadly, the first time it was iterated, using the function's output as its input, it immediately converged on a single value. Trying other initial values produced the same result. Reflecting upon this, Knuth wrote:

"...random numbers should not be generated with a method chosen at random. Some theory should be used." (Seminumerical Algorithms, Chapter 3.1).

_____________________________________________
Come to YAPC::Europe 2003 in Paris, 23-25 July 2003.