There’s a strange ( a.k.a interesting ) consequence of working with procedural buildings and … procedural content in general. In all cases, the hard part is to construct an algorithm that can produce something that looks plausible, which is much easier to say than to do, of course. However, if you do create a successful algorithm that can, for instance, create a manifold ( had to throw this constraint in, it’s easy to create disorder – we want order ) mesh of a building containing thousands of triangles, then what? Well since the building was created randomly the only thing required to store it are the boundary conditions that led to it’s construction. In this particular case, a rectangle outlining the boundary of the building and the seed on a random number generator. So compression wise, 4 numbers are all that are required to construct a building containing thousands of triangles, in terms of compression ratios … this is as good as it gets. Compression *is* entropy reduction. There’s nothing particularly new about this idea, I’m just propagating it, to create more order.
On another note, what about thinking? … that’s entropy reduction, it’s the process of trying to create order from disorder, therefore thinking is compression. Why not? 😉
On another note, think of Theoretical Physics, in this field they’re trying to do the same thing , their goal is to find compact mathematical descriptions of the governing rules of the universe. Once these rules are known all that is required is a simulator to execute the rules and the universe’s boundary conditions. Then, all that is you, is reduced to a mere 4 space time coordinates. Of course you could argue that the universe is not deterministic due to the uncertainty principle … I think this applies to measuring state, in the now, though, I don’t think it precludes the universe using a really cool pseudo random number generator at the beginning.
As per normal, more questions are raised than answers given, but it’s clear that entropy is one of our better ideas.