I don't have golly nor libGL on this desktop so I can't try out these rules. I'd be curious though as how you managed to run B0 rules at all given the whole-screen blinking and the deadly slow torus simulation? It'd be great if someone can convert the rule table to include a media state to prevent vaccum from blinking.Bullet51 wrote: The rule B3/S5 is somewhat less chaotic with a finite seed.
The rule B0/S8 features spontaneous string generation/annihilation with a finite seed, and the seed itself turned out to be unbreakable. Most other rules are chaotic, especially B1 and B2 ones. Nothing in B4 rules could escape their initial bounding box.
BTW, I managed to get some nice colored trajectory following Shalizi et al. 2006. I have to admit that working with ECA is a LOT simpler than with 2D-CA.
The algorithm can be found at this notebook (complex.ipynb) but it seems to identify oscillators as well as gliders, so the major utility will be background removal.
Further thoughts: these filters can only be applied in a post-hoc fashion. That is it does not guide the construction of glider, but only locate it after observing.