Markov Sort: an imprecise sorting algorithm

https://github.com/saniv/markov_sort

Learning 1-bit XOR requires just a few examples (no complex backpropagation crap and can be done online, during a playthrough). I'm really agreeing with MIT professor, who says that ANNs are not the only way to fit a curve (http://www.youtube.com/watch?v=q0pm3BrIUFo)


$ gcc markov.c && ./a.out
Training...
Input: 255,000
Guess: 255
Answer:255
Input: 000,000
Guess: 004
Answer:000
Input: 000,000
Guess: 004
Answer:000
Input: 000,255
Guess: 253
Answer:255
Input: 000,000
Guess: 004
Answer:000
Input: 255,255
Guess: 006
Answer:000
Input: 255,255
Guess: 006
Answer:000
Input: 000,000
Guess: 004
Answer:000
Input: 255,255
Guess: 006
Answer:000
Input: 255,000
Guess: 255
Answer:255
Input: 255,255
Guess: 006
Answer:000
Input: 255,255
Guess: 006
Answer:000
Input: 000,000
Guess: 004
Answer:000
Input: 255,255
Guess: 006
Answer:000
Input: 255,255
Guess: 006
Answer:000
Input: 255,000
Guess: 255
Answer:255
Input: 000,000
Guess: 004
Answer:000
Input: 255,255
Guess: 006
Answer:000
Input: 255,000
Guess: 255
Answer:255
Input: 255,000
Guess: 255
Answer:255
Input: 255,000
Guess: 255
Answer:255
Input: 000,000
Guess: 004
Answer:000
Input: 000,000
Guess: 004
Answer:000
Input: 255,000
Guess: 255
Answer:255
Input: 000,255
Guess: 253
Answer:255
Input: 000,000
Guess: 004
Answer:000
Input: 255,255
Guess: 006
Answer:000
Input: 000,255
Guess: 253
Answer:255
Input: 255,255
Guess: 006
Answer:000
Input: 000,000
Guess: 004
Answer:000
Input: 000,000
Guess: 004
Answer:000
Input: 000,255
Guess: 253
Answer:255
Input: 000,000
Guess: 004
Answer:000
Input: 000,255
Guess: 253
Answer:255
Input: 255,000
Guess: 255
Answer:255
Input: 255,255
Guess: 006
Answer:000
Input: 000,000
Guess: 004
Answer:000
Input: 000,255
Guess: 253
Answer:255
Input: 000,255
Guess: 253
Answer:255
Input: 000,255
Guess: 253
Answer:255
$