Fractals

A defining idea in fractal geometry is that detail can persist as an object is examined at different scales. Many mathematical fractals are exactly self-similar: smaller parts are scaled copies of a larger pattern. Natural shapes such as coastlines and branching forms may instead show statistical or approximate similarity over a limited range of scales. Fractal dimension describes how measured detail changes with scale and can be non-integer, unlike the familiar dimensions of a point, line or plane. For example, the boundary of the Koch snowflake has dimension log(4)/log(3), about 1.26: each step creates four smaller copies, each one-third the previous scale. This value describes scaling complexity, not an extra physical direction.

Many fractals are defined by repeating a simple rule, a process known as iteration. Repeatedly removing middle thirds produces the Cantor set; replacing parts of a line with smaller triangular bumps produces the Koch curve and, when three such curves are joined, the Koch snowflake. The Sierpiński gasket is formed by repeatedly removing smaller central triangles. Some fractals arise from iterating functions rather than altering a drawn shape. The Mandelbrot set, for instance, consists of complex parameters c for which the iteration zₙ₊₁ = zₙ² + c, starting at z₀ = 0, remains bounded. Computer visualization made the elaborate boundaries of such sets accessible at many levels of magnification.

The mathematical ideas behind fractals developed before the word itself. Nineteenth- and early-twentieth-century work on irregular curves and sets included examples by Karl Weierstrass, Georg Cantor and Helge von Koch; Felix Hausdorff introduced a generalized dimension in 1918. Benoît Mandelbrot later brought several strands together and coined the term “fractal” in 1975. His 1967 paper on the length of Britain's coastline discussed statistical self-similarity and fractional dimension as ways to describe the complexity of geographical curves. Thus fractal geometry grew from a longer mathematical history, while computers helped make many intricate examples easier to explore visually.

Fractal methods provide ways to describe or simulate rough, branching and scale-dependent patterns. They have been used in mathematical and scientific modeling, including studies of geographical forms and fluid behavior, and in computer graphics to generate irregular landscapes and other detailed images. Applying a fractal model does not mean that every natural object is an exact fractal: real measurements cover finite ranges, and natural self-similarity is usually approximate or statistical. Results therefore depend on the scale measured, the data available and the mathematical model chosen.