Nowadays computers are widely used to simulate real-life
processes and their mathematical models. These simulations require to
draw on the screen of a computer a finite set of points describing the
(often) continuous objects at a given precision. In this sense we need
to approximate points in metric spaces. We first review the complexity
approach to approximation theory for functional spaces which dates back
to Kolmogorov's work in the 50s-60s, then show the application of this
approach to fractal sets.