The formation of dynamical families is usually considered as the most spectacular outcome of catastrophic impact processes. However, several open problems involve the physical modeling and the interpretation of the observations. Most of the current ideas have been drastically revised on the basis of the first reliable hydrodynamical simulations, and the relevance of the Yarkovsky effect to shape the present distributions of the orbital elements has been put into evidence. However, open problems are still present. Some very recent findings seem to outline a somehow intermediate solution between the former paradigm and the later ideas.