In a recent paper, forthcoming in the Journal of Complex Systems, vol. 19, I present a method for studying the qualitative behavior of cellular automata and other abstract computing machines based on the approximation of their program-size complexity using a general lossless compression algorithm. I show that the compression-based approach classifies cellular automata (CA) into clusters according to their heuristic behavior, with these clusters showing a correspondence with Wolfram’s main classes of systemic behavior. I also present a Gray code-based numbering scheme for initial conditions optimal for this kind of investigations, and a compression based method for estimating a characteristic exponent in the form of a phase transition coefficient measuring the resiliency or sensitivity of a system to its initial conditions. I also conjecture that universal systems have large transition coefficients.