In the late 1950s and early 1960s, radio sky surveys were the centre of an intense and public debate-Big-Bang versus Steady-State cosmology-the arguments revolving about source counts and statistical interpretations in the face of instrumental complications. The 1965 discovery of the microwave background took the fire from the debate, but left the momentum in place for large-area radio surveys at different frequencies, and for extensive identification/redshift-measurement programs. By the 1970s the data enabled us to start disentangling the different populations of extragalactic radio sources. We could refine our taxonomy, and we could view the possibility of delineating individual cosmic histories and evolutions. We could at least describe a goal to elucidate the birth-life-death cycles of the objects involved [quasi-stellar objects (QSOs) and radio galaxies: together the 'active galactic nuclei' (AGNs)] whose unaccountably prodigious energies somehow produce the beautifully aligned radio structures with which we are now familiar. One part of John Bolton's vision to see how distorted a view of the AGN universe the original long-wavelength surveys provided. One legacy is thus the 'short-wavelength survey' for extragalactic radio sources, which has done so much to balance our picture of the radio sky. And indeed the legacy continues in the form of the immense sky surveys at present under way, complete with their sub-industries of radio-positioning and identification. From these, yet further results are emerging on spatial distribution and the skeleton structure of the universe. It is the purpose of this paper to outline something of this current view of the populations, their differences, similarities and unifying concepts.