Tutorials/Finding rules

From LifeWiki
Jump to navigation Jump to search

Welcome to the "big and scary" world of Other Cellular Automata! So, you want to create a nice rule. Really, there's no such thing as a "good rule" - different people are interested in different types of rules, but here are some general tips that should give you a general outline of what to do, because there are a lot of things you can do.

Outer-totalistic rules

Firstly, please be aware that because the amount of outer-totalistic rules is small (relative to the amount of Non-totalistic rules), most of the interesting rules here are known. But don't be frustrated! There are some recently (at the time of writing) found-to-be-interesting Outer-totalistic rules like B378/S2458 and B35/S2467, and there will probably be more.

For outer-totalistic rules, one of the most basic principles is a lack of B1 and B2. (B0 also tends to create strange explosions.) Both are guaranteed to make your rule explosive. B3 is also going to be a requirement, as it can be easily proven that without B0, B1 or B2, B3, no patterns can escape their bounding box and spaceships are not possible.[1]

Aside from that, it is generally a matter of removing transitions if the rule is too explosive, and adding them if the rule isn't chaotic enough. It helps to get a feel for what each transition is like, for example S0 and S1 tend to have very strong effects and S4 tends to create very large still lifes, in combination with other things.

Isotropic non-totalistic rules

Isotropic non-totalistic rules allow greater flexibility than Totalistic rules, as they allow you to specify the specific arrangement of cells. For a list of those arrangements of cells, see the article linked above in the title of this section.

Analogous to outer-totalistic rules, your transitions to avoid are B1e, B1c and B2a. You will also require one or both of B3i and B2c to make patterns able to escape their bounding box, and one or both of B2e and B3a to enable patterns to escape their bounding diamonds.

Making isotropic non-totalistic rules can be difficult, however once you get a "feel" for it you will start to notice general trends in rule "characters", such as the typical behavior of rules with S1 and a lot of B2 transitions.

It helps to start first with a well known "good" rule, but otherwise there isn't much else to say. It may also help to think up a spaceship, attempt to make it work and then go from there, adding/removing transitions as required. Or, you could start with a small pattern, say, a pre-traffic light (one half of a pre-pulsar) and try to make a rule where it does something nice, such as become a large spark.

Further reading