Computer simulations have been characterized by making a sharp contrast with others scientific practices. In this paper we propose to understand computer simulations from interventive practices in a particular field: life sciences. We defend two theses. First, to characterize computer simulations requires to understand experimental practices. Second, we argue that to understand experiments requires a previous notion of simulation. In other words, we defend that the practices associated with simulations and experimentation are two complementary aspects, at least in life sciences.