Since I was a kid I've wondered about something like this, although my (then) 12-year old way of expressing the question was a lot different (and a lot simpler in scope). What if you wrote a simple program that filled a 1000x1000 pixel matrix of 24-bit RGB values with every possible combination? The program could be short, but really the language is irrelevant, as you could design a special language with one argumentless 'instruction' that did only this. (We could start now getting into whether the language's host environment (OS) can be regarded as simply a fancy immediate-mode interpreter, but thats a different topic)
What are the implications? Forget how long it would take or the fact that perhaps 99% of the resulting images would be apparent garbage (could be looking at every square meter of sidewalk on earth), isn't it possible that every conceivable image in the entire universe would eventually get drawn? But how can that be, since although the number of possible combinations (64^1000000) is unfathomably large it is still finite, but isnt the universe (and therefore the subset of the universe that is visible in 24-bit RGB) infinite? Or is it? Every person's face that ever lived, every rock, stone, every animal, evey bacterium, every star, - everything, from every possible angle, would eventually appear.
"Aha", you say, "but at that fixed resolution two grains of sand/two stars/two ashley twins might look identical, pixel-for-pixel!" Right, but remember that all possible distances from each are shown, so if you multiply the zoom by two and tile 4 adjacent 1000^2 pixel zoomed images into a square the differences can be appear; if not, do it again and again until the sufficient resolution is obtained to show the differences.
Now remember my 1000x1000 grid is arbitrary, you can use a 10x10 grid if you like, and in fact that makes this more troublesome because the pace is much smaller: there are vastly fewer 10x10 24 bit images, but the zoom-and-tile method can still be applied - so does this mean the whole universe can be shown in 24^100 tiles? What about a 5x5 grid? 2x2? Help me out here because I'm trying to figure out what's missing in my logic - there's no way the limited number of permutations of 2x2 pixel grids is sufficient to express every image in the universe, but by zooming and tiling it seems like its possible. Or not?
Rob Cebollero (using an old account since I am not near my usual machine)
- Opinions subject to change without notice. -RC