Twenty-one years past since Weiser's vision of ubiquitous computing (UbiComp) has been written, and it is yet to be fully fulfilled despite of almost all the needed technologies already available. Still, the widespread interest in UbiComp and the results in some of its fields pose a question: why we are not there yet? It seems we miss the 'octopus' head. In this paper, we will try to depict the reasons why we are not there yet, from three different points of view: interaction media, device integration and applications.