A unit cell (or metacell) is a subset (usually rectangular or square) of the Life plane that tiles over the plane, along with a fixed number of distinct patterns, with each tile assuming one of the patterns, such that it simulates a cellular automaton, possibly itself. A unit Life cell is a unit cell that simulates the Game of Life. To avoid single cells themselves being considered unit cells, the size of a unit cell must be greater than 1×1. It is also a restriction that only finite sized patterns are accepted as unit cells excluding infinite one-cell thick bars in HighLife for example, which simulates Rule 54; many amphichiral 1D elementary automata can be embedded in other rules via this method.
Simulating a 2D cellular automaton
The first unit Life cell was constructed by David Bell in 1996. It employs standard glider logic to determine whether or not a glider should be present. The two states differ by a single glider. In 2004, Jared Prince modified David Bell's unit Life cell to support two (and therefore multiple) layers of Life universes, coined "deep cell".
Around 2005, the OTCA metapixel was constructed, which can simulate any Life-like cellular automaton. Designed to run quickly in HashLife, it has the advantage of having two states that are clearly distinct when zoomed out.
The P1 megacell, designed in 2008, has clearly visible states like the OTCA metapixel. It is capable of simulating any rule, including non-totalistic and asymmetric rules, that uses the standard eight-cell neighborhood. It also has unusual positioning, being a square with diagonal edges. This allows much of its information to be transmitted with gliders.
The 0E0P metacell is currently the largest unit cell. Unlike other unit cells, it can function without any form of support circuitry, allowing it to produce the first examples of exotic patterns such as parity-rule replicators and reflectorless rotating oscillators in Life when programmed correctly. Rather than having an "on" state or an "off" state, it deconstructs itself when it turns off and is built by its neighbors when it is turned on.
Simulating a 1D cellular automaton
Complex one-dimensional cellular automata usually can be simulated via constructed unit cells. Purpose of this is basically the containing automaton inherits some useful properties of the embedded, which is usually not trivial to prove. Such properties are logic universality, Turing-completeness etc. It is known that many examples are constructed in order to simulate Rule 110 and allegedly Rule 30 in one case. One is constructed in CGoL by Jason Summers and later it was trivially shown using Golly that it is a polyglot (works in EightLife, too)
Selection of life-like rules with constructed W110 unit cells
- B3/S23 polyglot: works in Life, EightLife, Pedestrian Life and HoneyLife
- Banks-I: proves logic universality using Cook's results, simplifying Banks' proofs from 1971
- DryLife and 3 more
- Paul Callahan (March 1, 1996). "The Unit Life Cell".
- Jared Prince (September 27, 2004). "Game of Life Deep Cell".
- Brice Due (May 28, 2006). "How Does It Work?". OTCAmetapixel. Retrieved on March 26, 2009.
- Hienrich Koenig (December 21, 2005). ""Rule 110" Unit Cell". Game of Life News. Retrieved on March 29, 2018.
- Peter Naszvadi (November 21, 2017). FWSS-less MWSS-track-only Rule-110 Unit Cell (discussion thread) at the ConwayLife.com forums
- Peter Naszvadi (November 1, 2017). Re: List of the Turing-complete totalistic life-like CA (discussion thread) at the ConwayLife.com forums
- Peter Naszvadi (October 14, 2017). Re: List of the Turing-complete totalistic life-like CA (discussion thread) at the ConwayLife.com forums
- Peter Naszvadi (March 29, 2018). Re: List of the Turing-complete totalistic life-like CA (discussion thread) at the ConwayLife.com forums