Diagram of information complexity, copied from the text quoted below
Most of the proposed measures [of information complexity] differ from each other but share at least one important characteristic, in that strictly regular things as well as strictly irregular things are ‘simple’, while things that are neither regular nor irregular are ‘complex’… On one extreme we have the pure regularity found in Euclidean objects. Deterministic fractals, while not simple, have compact algorithmic descriptions but elaborate structures. Fractals fail to be strictly simple because even though they are defined by short programs, these programs must be made to run forever in order to completely express the infinite self-similarity that they contain. On the other extreme is pure noise. Brownian motion is slightly more complex than pure noise because Brownian processes must have ‘memory’, in that every random injection is always made relative to the previous state; thus, a random walker’s position is not just a random location but the sum of an infinite number of random steps. Hence, fractals, both deterministic and stochastic, while not simple, seem to be on the edge of being complex.
From: “The Computational Beauty of Nature” by Gary William Flake, pg 135.