The most extensive part of the third revision was precipitated by the article
H. Chaté and P. Manneville
Evidence of Collective Behavior in Cellular Automata,
Europhysics Letters 14 409-413 (1991),
and several events following its publication. The principal item of interest is a cyclic overall behavior of the probability densities as one of these automata evolves; the automata themselves are multidimensional.
Automata of different dimensionalities can have neighborhoods of the same size, depending on the number of ways that the size can be factored. Such differences have to affect the correlations between different cells of the neighborhood when calculating probabilities, providing a motivation for using different neighborhoods within the same program. Their adjunction of a program set emphasizing one-dimensional automata was due to the elaborate facilities for displaying probabilities which already existed.
Harold V. McIntosh