In 1967, Benoit Mandelbrot publishes in Science a paper with a deliberately provocative title: "How Long Is the Coast of Britain?". The answer, Mandelbrot shows by reviving a forgotten observation by meteorologist Lewis Fry Richardson, is that the question lacks a fixed meaning: the smaller the measuring stick used, the longer the coastline turns out to be, because each straight stretch reveals, at higher resolution, additional sinuosity that repeats at ever smaller scales. Mandelbrot formalizes this property — self-similarity across scales — through the concept of fractional dimension: while a straight line has dimension 1 and a plane dimension 2, an irregular coastline can have an intermediate, non-integer dimension that objectively quantifies its degree of roughness. In 1975 he coins the term "fractal", from the Latin "fractus" (broken, fragmented), to name this class of geometric objects, and develops the theory in technical detail in "Les Objets Fractals" (1975) and, more visually and accessibly, in "The Fractal Geometry of Nature" (1982), lavishly illustrated with the first computer-generated images of these structures. Mandelbrot shows that natural phenomena that Euclidean geometry described very poorly — coastlines, mountains, clouds, river networks, tree branching, blood vessels — all exhibit this same property of fractional self-similarity. The field quickly transcends pure geometry: fractals become a modeling tool in chaos physics, materials science, biology, seismology and, notably, in Mandelbrot's own work, in the analysis of financial markets, where he had already applied similar ideas in 1963 to the study of cotton price fluctuations.