In this short review we summarize the outcome of the large amount of effort made during the past decade from both the experimental and the theoretical point of view in order to understand the effect of blending on the segmental dynamics in polymers. Each of the two families of models proposed-one based on thermally activated concentration fluctuations, the other on chain connectivity effects-account for each of the two main experimental observations: the broadening of the component response with respect to that of the homopolymer and the dynamic heterogeneity, respectively. The complementarity of these approaches, their main achievements and failures, are critically revised. We also include recent results on blends of components with very different mobilities. In the neighbourhood of the glass-transition of the slow polymer, the dynamics of the other component seem to be confined within the frozen chains. We suggest possible ingredients and new routes to be considered in order to elaborate more predictive theoretical frameworks for all these phenomena.