# Glossary

**Cellular automaton / CA**  
A discrete dynamical system comprising cells having states that are updated according to a local transition rule or evolution function.

**Evolution function `F`**  
The protected local transition function applied to a cell and a bounded neighborhood to compute a next state.

**Cell state**  
The state value stored for a CA cell. In the current contemplated embodiment, each cell may be represented by a compact multi-bit state, for example a 6-bit state.

**State image / image program**  
An initial, maintained, or loaded spatial configuration of CA cell states that performs computation by evolving under `F`.

**Memory-image-defined computation**  
Computation in which the effective machine or program is encoded in a CA state image rather than in a conventional sequential instruction stream.

**Persistent propagating structure**  
A localized CA structure, such as a glider, that propagates through the lattice while preserving recognizable identity or information.

**Interaction region**  
A spatial region of a CA image configured so that incoming propagating structures interact to perform logic, routing, detection, emission, transformation, or simulation behavior.

**Output region**  
A predetermined memory region read after one or more CA update cycles to obtain computational output.

**Effective hardware architecture**  
The logical processor, accelerator, simulation engine, or computation fabric defined by a loaded CA image, as distinguished from the fixed physical update circuitry.

**Pull-based update**  
A CA update method in which each cell computes its own next state from its current state and neighboring states, thereby avoiding multi-writer conflicts.

**Symmetry-preserving conflict resolution**  
Conflict handling in which simultaneous neighbor influences are resolved by a rule that is equivariant under selected lattice rotations, reflections, or coordinate transformations.

**Bias-compensated deterministic update**  
Conflict handling in which deterministic local priority orientations vary across cells, rows, planes, tiles, or cores according to a balanced pattern so that no lattice direction is globally preferred.