I also think they are one and the same. Because a network is a markov chain, with a huge, lossily compressed, state transition table. (that the chains weren't d

I also think they are one and the same. Because a network is a markov chain, with a huge, lossily compressed, state transition table. (that the chains weren't designed by a human is irrelevant to what they are). given that, you can very mechanically convert it to a statistical compressor by just attaching an entropy coder. This gets very good compression.