Iteration and recursion
Fractals often begin with a short instruction repeated many times. The richness comes from feedback: each result becomes the starting point for the next step.
Two ways to repeat
Iteration applies the same operation to a changing state: z becomes f(z), then f(f(z)). Recursion defines an object in terms of smaller versions of itself. A program may implement a recursive geometry iteratively, so the mathematical idea and the coding technique should not be confused.
The difference becomes practical when a construction is programmed. A recursive description of the Sierpiński triangle may call a function for three smaller triangles; an iterative implementation can process the same states in a loop or queue. The visible result can match while memory use and execution order differ. In dynamical systems, iteration usually denotes a temporal sequence of states as well. Mathematical and software-engineering meanings should therefore be kept deliberately separate.
Finite stages and infinite limits
Every displayed Koch curve or Sierpiński triangle is a finite approximation. The mathematical object is the limit approached as construction steps continue. Pixel grids make later stages indistinguishable, even though the abstract definition has no last step.
Intermediate stages are not inferior versions of the limiting object; they are the only versions a display or material can realize. Their depth determines which structures remain visible and which collapse below one pixel. For geometric constructions, the required stage can be estimated from the smallest representable edge. In escape-time fractals, the iteration cap plays a related but different role: it limits observation time rather than geometric construction depth.
Feedback creates sensitivity
In dynamical fractals, a small change in the initial value or parameter can send an orbit toward a different outcome. Boundaries record those competing futures. This is why a formula as compact as z² + c can produce a frontier with seemingly inexhaustible structure.
Feedback amplifies differences only where the dynamics is sensitive. Other initial states may converge rapidly or settle into stable cycles. The mixture of stretching, folding and capture is what creates a rich phase space. A good experiment therefore changes more than one starting value: it follows neighborhoods, discards a transient period and checks whether the observed behavior remains numerically stable. A single spectacular orbit rarely tells the whole story.
Sources and further reading
This article summarizes the following specialist sources in original wording. Accessed and editorially reviewed 12 August 2026.
- Fractal Geometry: Mathematical Foundations and ApplicationsWiley
- FractalWolfram MathWorld
Continue the idea
Complex numbers as a plane
A visual introduction to complex numbers, magnitude, angle and the squaring operation behind Mandelbrot and Julia images.
The Mandelbrot set
What the Mandelbrot set contains, how z² + c creates its boundary and why deep zoom reveals related rather than identical copies.
Iterated function systems
How a small collection of contracting transformations defines a unique fractal attractor.