would be an abstract answer (because it would not immediately link a formal model with the biological reality that inspired it or with possible implementations) from the other hand it would be a deep answer. This answer would allow us to understand more fundamental features that have to be present in a formal system in order to compute, it would allow us to classify different formal systems in a uniform way, it would possibly give us new tools to prove the computational power and other properties of different kinds of systems, etc. Recently a way to answer this why, using the topology of information flow, has been suggested [3, 5, 6, 7] In this paper we survey the links between topology of information flow and conformon P systems. At the same time we show how a similar answer could be given for other formal models of computations (P systems and not). The given directions of research and open problems are meant to inspire further developments in this line of research. References