For decades radio surveys have been a leading agent for extragalactic research, as testified by the breakthroughs they triggered, from the discovery of cosmic evolution, to quasars and the first high-z galaxies. They continue at the forefront of astrophysics and cosmology; e.g., via large-scale structure studies they pose challenges for buildup of the cosmic web; and via downsizing and AGN feedback deemed to produce this downsizing, they have come to the fore in modelling galaxy formation and evolution.
But our physical understanding of the origin and evolution of the AGN-powered radio emission is still poor. While physical models for the cosmological evolution of galaxies and radio-quiet quasars have been progressing rapidly in recent years, the main progress on the radio side has been towards a phenomenological description of evolution of various radio AGN types.
Even on the phenomenological side, there are aspects that are not fully understood. The epoch-dependent luminosity functions of galaxies and radio-quiet quasars are now quite accurately determined up to high redshifts and there are attempts to provide physical explanations for the evidence for earlier formation of the more massive objects. Direct evidence for a substantial decline of the space densities of radio AGNs at z 2 remains somewhat controversial, although modern evolutionary models accounting for the observed counts and redshift distributions do include such a decline.
The origin of these uncertainties remains as it has been for the last 30 years - a lack of concerted effort to obtain complete redshift sets for radio AGN samples. The new redshift surveys (2dF, SDSS) have helped greatly in defining local space densities. But beyond two or three of the brightest samples such as 3CRR, there are no samples with complete redshift data. Even samples of 100 to 200 objects would suffice for most purposes, and could be easily obtained with 8- to 10m-class telescopes. Our deficiency in this regard is highlighted by the reliance of most analyses needing complete redshift information on the Dunlop & Peacock (1990) model distributions - even 20 years on. Doubtless the planned new deep optical and infrared wide-field surveys such as PAN-STARRS (Hodapp et al. 2004) and those with the VST and VISTA (Arnaboldi et al. 2007) will help. However there will remain a need for individual pursuit of the faintest members in samples via deep imaging in different optical and IR bands and via fast spectrographs to complete the redshift information in samples of limited size.
Sixty years since their discovery, radio AGNs remain at the forefront of astrophysics and cosmology. Our continued attempts to solve the mysteries which still surround them will doubtless lead to fresh discoveries of impact as great as those which have distinguished the first 60 years.
We are grateful to the referee, Jim Condon, for a very careful reading of the manuscript and for many very useful comments and suggestions. Thanks are due to Benedetta Ciardi for having provided the 1.4 GHz counts of GRB afterglows in tabular form. GDZ and MM acknowledge partial financial support from ASI contracts I/016/07/0 "COFIS" and "Planck LFI Activity of Phase E2". JVW acknowledges support via Canadian NSERC Discovery Grants.