Some decades ago, the enthusiasts for computational modelling predicted the demise of experiments and their replacement by computational models. This was predicted not only for turbulent single phase flows (as exemplified by the “numerical wind tunnel” concept) but also for the more difficult case of multiphase flows where deformable interfaces coupled with turbulence present a formidable challenge. In the event, these predictions of doom for experiments have proved somewhat premature. There have been remarkable advances in computational methods but there are still limitations in our ability to predict turbulent flows at Reynolds numbers of real engineering interest and to predict real multiphase flows. Coupled with these limitations on computing, one has to recognise that the experimentalists have got smarter, introducing new advanced techniques which allow fresh insights into the phenomena occurring. In reality, both advanced experiments and state-of-the-art computing are simply alternative and complementary tools in exploring the physical world in general and multiphase systems in particular.