the Koch snowflake
/ KOKH (rhymes with 'loch') /
Here is a shape that seems to break the rules: a closed curve enclosing a finite, ordinary patch of area, yet whose boundary has INFINITE length. The Koch snowflake is the classic example, a jagged crystalline outline you can draw by an endlessly repeated rule. It looks like a snowflake, and it became one of the first proofs that a curve can be continuous everywhere yet so crinkled it has no tangent line anywhere.
You build it by a simple recipe applied forever. Start with an equilateral triangle. Now take each straight side, divide it into three equal parts, and replace the middle part with the two other sides of a smaller equilateral triangle pointing outward — so a straight segment becomes a four-segment zigzag with a little spike. Then do exactly that to every one of the new, shorter segments, and keep going without end. Each round multiplies the number of segments by 4 while shrinking each to one-third the length, so the total perimeter is multiplied by 4/3 every step. Repeated forever, the perimeter grows past any bound: it is infinite. Yet the whole figure stays trapped inside a small circle, so the area it encloses is finite (it works out to 8/5 of the starting triangle's area). The boundary is self-similar — each little bump is a scaled copy of the whole edge — and its fractal dimension is log 4 / log 3, about 1.26: more than a 1-dimensional line, less than a 2-dimensional region.
The Koch snowflake, introduced by Helge von Koch in 1904, was a deliberate 'monster' that forced mathematicians to sharpen what they meant by length, dimension, and curve. Its lesson echoes in the real world through the coastline paradox: measure a coastline with a finer ruler and you always find more length, because real coasts are crinkly like Koch curves, so the 'length of a coastline' has no single well-defined answer. One honest clarification: infinite perimeter with finite area is not a paradox or a contradiction — it is simply what happens when a boundary is infinitely rough; perimeter and area measure different things, and there is no law forcing a finite area to have a finite boundary.
Track the perimeter. Start with a triangle whose side is length 1, so the perimeter is 3. After one step each side becomes 4 segments of length 1/3, so each side's length is 4/3 and the perimeter is 3 times 4/3 = 4. After two steps it is 3 times (4/3)^2 = 16/3, about 5.33. Each step multiplies by 4/3, so after n steps the perimeter is 3 times (4/3)^n, which races off to infinity as n grows — even as the enclosed area settles down to a finite limit.
Perimeter multiplied by 4/3 each step diverges to infinity, while the area converges to a finite value.
Infinite perimeter enclosing finite area is not a contradiction. Length and area measure different things, and an infinitely crinkled boundary can be arbitrarily long while the region it bounds stays small — exactly the coastline paradox in miniature.