is a very simple CellularAutomaton
invented by JohnHortonConway
and popularized by the writings of MartinGardner
(e.g., Scientific American 223 (October 1970): 120-123). GOL is capable of universal computation and of self-reproduction, but is vastly simpler than the first CellularAutomaton
capable of self-reproduction, discovered in the 1940s by JohnVonNeumann
A quest was launched to find programs (configurations) with interesting properties. For many it was a first taste of non-VonNeumannArchitecture
What fascinates me is that the patterns that do something useful are often very decorative. They look at first sight as if they may contain a lot of AlexandrianCenters. Does anyone else share this impression? -- KeithBraithwaite
s offer modern kids another alternative.
is a CellularAutomaton
, a classic piece of theoretical computer science, the foundation of JohnVonNeumann
's self-reproducing system (so CellularAutomata
are not in fact non-VonNeumannArchitecture
!), and a good example of ArtificialLife
. -- BillTozier
It is possible to build Life modules that act as logical AND, OR and NOT gates, therefore a TuringMachine
, and therefore any computable model (including self-awareness if self-awareness is computable). (True, but with all the other interesting capabilities of the game, is seems a shame to "reduce" it a mere Turing computing engine. Seems more more interesting to discover NonTuringComputing with it.)
Oh, I beg to differ! Isn't this the inverse
of "reducing" it? It seems to me that this demonstration proves that it is larger
, and therefore more interesting.
The solution to this paradox is that ReductionIsGood?. It sounds like a negative term, but it isn't. For me, reduction is almost synonymous with understanding.
This is also a graphic demonstration that emergent behavior at one abstraction level can be used to construct deterministic systems at another. This puts to bed, in my view, the logical fallacy that underlies many arguments against artificial life and artificial consciousness. The observed presence of "order" at one abstraction level (such as when looking at the elements of a computing engine built from Life) does not imply intentional design at lower abstraction level (such as the rules of the environment in which those elements operate).
In late 2000, a TuringMachine
has been demonstrated in a 1714 x 1647 GameOfLife
pattern implementing a 3-state 3-symbol machine which takes 11040 generations for a single cycle. Pictures and explanations at:
The author says: "The design for this Turing Machine is extendible by expanding the size of the Finite State Machine part and storing different numbers in the memory cells. The maximum size is 16 states and 8 symbols. This is sufficient for a Universal Turing Machine."
That thing looks amazingly like models of DNA transcription!!!
was one of the seeds of StephenWolfram
's search. He states everything can be explained through simple self-replicating rules. He wrote a gigantic book about it, A NewKindOfScience
. Now published online: http://www.wolframscience.com/
There is a ForthLanguage
implementation of the GameOfLife
In a different sense: RealLife GameOfLife
In the context of your own life, the GameOfLife
has to be played on a field where LifeIsaBigMessyGraph
, the GameOfLife
is likened to be engaged in "Concurrent Attention-Seeking Games": GameOfChess
, etc all in time limited fashion. The involvement alone stops people from exploring MeaningOfLife
in a meaningful manner.