Talk:Search and Representation
From Maze
Explanation of binary number. Good. In talking about randomness, and comparing the uniform distribution view of randomness with the algorithmic complexity theory view of randomness (Chaitin, Kolmogorov, etc.), I was planning on talking about binary sequences, so we'll need to introduce binary numbers somewhere in the book. As a pet peeve, in introducing binary numbers, we'll need to be careful to avoid the simplistic binary essentialism that plagues so much writing on "the digital" (which is so much more productively conceived of as "the computational").
The section on combinatorics directly connects with notions of randomness. Visually recognizable patterns are less "random" under algorithmic complexity theory, which connects well with our intuition of random (everyday intuition of the random has more to do with detection of patterns than with uniform distributions). Also, certain patterns (e.g. a screen full of all 1s or 0s) are literally impossible given a specific pseudo-random number generator - many (most? all?) specific prngs don't actually sample from the entire 2^(24 * 40) = 2^960 possible patterns and so are not "random" for the uniform distribution sense of the word. I'd like to argue that in the context of expressive processing, we don't really care about the uniform distribution notion of randomness, but rather the algorithmic complexity theory notion (a sequence is random if the length of the smallest program that can produce the sequence is equal to the length of the sequence). Given a god's eye view of 10-print, the sequences are obviously far from random, since a very small program is producing them. But, given our limited human intelligence, we are not able to discern the short program producing the sequence, and thus perceive the generated sequences as irreducible sequences of (aesthetic) choices.
The section on how to interpret the pattern as a maze directly overlaps with maze movement I was going to talk about. It looks like Warren is using a different set of movement rules than the path-like rules (I really need a diagram to help figure this out). This is great. I was hoping to talk about different movement interpretation rules as well in the context of how different movement rules effect our perception of movement. Using the path-like movement rules (which my walkers do) can make it harder to perceive the walker as moving between the walls - it looks more like it's walking on the walls. Different movement rules change the interpretation of location within the maze. The perception of location is problematic in the 10-print pattern because there is no floor, only walls.
Seventh section of evaluation of pattern as potential maze. Great. Has obvious parallels and overlaps with the generate-and-test maze generator I wrote. Warren, would love to work on this with you.
Eighth section. Nice. I was planning on writing a basic program to color unique paths, but hadn't made the connection to the four color theorem.
