Self-similarity is a striking and ubiquitous characteristic of structures in geology. On the other hand, capturing of scale effects is one of the main challenges in geotechnical engineering. There, the usual approach is to extrapo-late laboratory findings toward large scale problems. Whereas this procedure is passable for small grained soils, it proves unrealistic for coarse grained soils and rock, especially jointed one. As a consequence, the prediction of e.g. water inrush into a large tunnel based on the examination of rock samples and spot-wise packer tests proves to be hardly possible. The new branch of fractal analysis encompasses both notions, self-similarity and scale effects. Therefore, it pricks up the geotechnical engineer's ears and evokes the hope that it can help to treat the related problems. A more thorough view, however, shows that the exploitation of fractal analysis in the above sense is not so straightforward. First of all, our limited range of observance over several scales inhibits to check and quantify the scale effect. At that, we lack the mathematical tools to treat problems that exhibit a continuous scale dependence, as this is the case with fractal materials.