We present a method that allows us to produce aligned metallic nanowires on almost any material. This method recently enabled us to produce interconnected platinum nanowire networks on polymeric Nafion® or gold nanowires on Teflon AF. We will explain the principle of this method by means of a reference material, the so called layered transition metal dichalcogenides (TMDC) crystals. The method is based on producing cracks in thin films as templates, using large sticking coefficient differences and long diffusion lengths. The underlying mechanism that forms the template cracks is mechanical stress. In the case of the TMDC crystals this stress is introduced in the substrate by an electronic interaction between the metal used for the nanowires and the TMDC-crystal. The large difference in the condensation coefficient is an immanent property of layered crystals for many different adsorbates. Obtained by cleavage along a so called “Van-der-Waals-gap”, surfaces of such crystals are atomically flat over hundreds of microns. As a consequence of their layered character the surfaces have no reconstruction and almost no step edges or other defects, which is the basis for the long diffusion length. In contrast, the rare defects have a very high sticking coefficient and act as excellent nucleation centers. Therefore, after metal evaporation on such surfaces in UHV, various structures can be formed in a self organized processes. We observed, e.g., clusters in fractal or geometric arrangements or large nanowire networks on the surfaces [1]. In order to understand, why the different structures form, a systematic study of the growth parameters, (nucleation, diffusion length, evaporated metal, influence of the substrate-crystal), is necessary. Therefore, we first carried out diffusion studies on these surfaces. We could show that in extreme cases (Cu on metallic TaS2) diffusion length of more than 50νm could be observed combined with a nucleation probability of almost zero. This is evident from a growth mode showing similarities with the DLA (Diffusion Limited Aggregation) growth process. In contrast, metal diffusion (Cu) on the geometrically similar surfaces on the semiconducting (WSe2) surface shows much a shorter diffusion length and no DLA growth. We suggest a model to explain the different diffusion behavior as a key to understand the different self organized structures. Learning from the TMDC crystals, we show the application on technological more important materials.