We have addressed the central question in two-dimensional disordered systems: whether all states are localized even at infinitesimal disorder, as has been believed for over two decades but has come in question after recent experiments. We argue that the answer could still be yes since above a certain density the electrons can be treated effectively as non-interacting. Below this density, though the interactions among electrons get weaker they become more and more important as the electron density reduces. This can cause delocalization and produce a metallic phase. Some recent experimental results have been analysed to understand the nature of the newly found metallic phase at low electron-densities, and also the insulating phase at lower densities.