A measure of the algorithmic information content of a finite object, defined as the length of the shortest description (program plus input) on a fixed universal description machine that produces the object.

Principle

The complexity of an individual finite string is the minimal length, up to an additive constant dependent on the chosen universal machine, of a program that outputs that string and halts.

Demonstration

A periodic binary string such as '010101...0101' has low Kolmogorov complexity because a short program can generate the repetition, whereas a long string produced by fair coin flips typically has near-maximal complexity.

Misapplication

Treating the quantity as computable for arbitrary inputs, or comparing raw values across different unspecified universal machines without accounting for additive constants, leads to invalid conclusions.

Consequence

Provides a formal basis for individual randomness, noncompressibility lower bounds, and theoretical limits on algorithmic data compression and inference.

Reversal

Shifting to ensemble-based measures (such as average-case information across a distribution) inverts focus from single-object descriptive minimality to statistical uncertainty across samples.

Boundary

Defined for finite discrete objects relative to a choice of universal description method; it is not computable in general and depends on the invariance up to a machine-dependent additive constant.

Semantic Tension

Contrasts with Shannon-style information which quantifies average uncertainty in a stochastic source; Kolmogorov complexity quantifies individual description length independent of probabilistic ensembles.

Synthesis

Kolmogorov complexity formalizes the shortest effective description length of a concrete object relative to a universal description scheme, grounding algorithmic randomness and compressibility as properties of individual instances.