On Wed, 8 May 2019 at 03:39, Richard Damon <rich...@damon-family.org> wrote: > My experience is that the wrap around is common, as otherwise the hard > edge causes a discontinuity in the rules at the edge, so any pattern > that reaches the edge no longer has a valid result. The torus effect > still perturbs the result, but that perturbation is effectively that the > universe was tiled with an infinite grid of the starting pattern, so > represents a possible universe.
In my experience, "simple" implementations that use a fixed array often wrap around because the inaccuracies (compared to the correct infinite-area result) are less disruptive for simple examples. But more full-featured implementations that I've seen don't have a fixed size. I assume they don't use a simple array as their data model, but rather use something more complex, probably something that's O(number of live cells) rather than something that's O(maximum co-ordinate value ** 2). Paul -- https://mail.python.org/mailman/listinfo/python-list
