I review our ideas on how massive compact galaxies can form and evolve. I will show that we do not fully understand how massive and passive galaxies (M>3x10^11 M_sun) as compact as the galaxies we see at redshift z>2 assemble their stellar body. I will argue that their morphology, as reveled by HST images, is telling us that merging of pre-existing stellar progenitors cannot be a viable mechanism for this class of galaxies. Rather, their compactness and high stellar density are the telltale of a highly dissipative gaseous process, whose details, however, still elude us. I will also present plausible potential progenitors, i.e. star-forming galaxies at redshift z>2, and discuss that their properties are consistent with stellar feedback being the agent of a rapid quenching of the star formation activity. These galaxies are very compact systems that evolve at essentially constant size and quench before major merging event transforms their morphology. They might be the descendant of the compact disks observed at z>6.