For some time the theory of cellular automata consisted of the inevitability of evolution into cycles, the existence of pristine initial configurations, and a respectful reverence for the complexity of von Neumann's detailed plans and the philosophical implications of his efforts. In the practical direction, a variety of efforts with image processing and photographic interpretation were understood to be related as much to cellular automata as to Fourier or Walsh transforms[58], and some thought was given to designing circuit arrays as automata.
One way or another. the lore of automata theory has gradually spread around. Studying cellular automata has become a popular computer game in addition to being a learned academic subject in several stages. Perhaps the initial step resulted from John Conway having decided to review the area just at the time that computers acquired versatile visual display equipment.
By taking the approach of first defining the automaton, then examining its capabilities, Conway[28],[29] encountered a two-dimensional binary cellular automaton of incredible elegance and beauty. Christened Life , it was eventually shown to have the same powers of universal computation and self-reproduction as von Neumann's automaton[9]. Even so, much work and several years were required to acquire familiarity with Life ; many of the results were reported as they were being discovered in a quarterly newsletter published for nearly three years by Robert T. Wainwright[68]. Later some of the information was published more formally in a book written by William Poundstone[57]. The substance of all the notices in Martin Gardner's columns was collected in one of his reprint volumes[29].
From the outset Wainwright undertook to catalog the Life configurations that were being discovered, classifying them roughly along the lines of the general theory. Conway knew about some travelling configurations called gliders, but was apparently unprepared for the discovery of first ``glider guns'' and then ``puffer trains'' which between them permitted the assembly of incredibly complex artifacts. Nevertheless he was able to use them to achieve von Neumann's results, although in terms of an equally arduous construction spreading over a fantastic area if it were ever to be realized in practice. The number of contributors to Wainwright's newsletter was considerable, but the contributions of William Gosper and several associates at MIT's Artificial Intelligence Laboratory were fairly outstanding. Some of their adventures have been recounted in Stephen Levy's book, The Hackers[48].
Conway's choice of a particular rule and lattice resulted from careful experimentation, giving Life a relatively distinguished setting. Poundstone[57] reports experiments with an alternative rule, Packard and Wolfram[55] surveyed numerous two dimensional automata; in both cases reporting that their automata fail to meet Conway's criteria. It would appear to be an interesting challenge, either to account for the singular nature of Life or to encounter additional specimens.
Besides offering a new and interesting automaton to the existing repertoire, Conway's contribution was to begin with specific automata with the intention of ascertaining their potential behavior, rather than beginning with the application and searching for the automaton which would fulfil the requirement.