The CMB is a mature subject. The spectral distortions are well understood, and the Sunyaev-Zeldovich effect provides a unique tool for studying galaxy clusters at high redshift. Global distortions will eventually be found, most likely first at very large l due to the cumulative contributions from hot gas heated by radio galaxies, AGN, and galaxy groups and clusters. For gas at ~ 106 - 107 K, appropriate to gas in galaxy potential wells, the thermal and kinematic contributions are likely to be comparable.
CMB anisotropies are a rapidly developing field, since the 1992 discovery with the COBE DMR of large angular scale temperature fluctuations. At the time of writing, the first acoustic peak is being mapped with unprecedented precision that will enable definitive estimates to be made of the curvature parameter. More information will come with all-sky surveys to higher resolution (MAP in 2000, PLANCK in 2007) that will enable most of the cosmological parameters to be derived to better than a few percent precision if the adiabatic CDM paradigm proves correct. Degeneracies remain in CMB parameter extraction, specifically between 0, b and , but these can be removed via large-scale structure observations, which effectively constrain via weak lensing. The goal of studying reionization will be met by the interferometric surveys at very high resolution (l ~ 103 - 104).
Polarization presents the ultimate challenge, because the foregrounds are poorly known. Experiments are underway to measure polarization at the 10 percent level, expected on degree scales in the most optimistic models. However one has to measure polarisation at the 1 percent level to definitively study the ionization history and early tensor mode generation in the universe, and this may only be possible with long duration balloon or space experiments.
CMB anisotropies are a powerful probe of the early universe. Not only can one hope to extract the cosmological parameters, but one should be able to measure the primordial power spectrum of density fluctuations laid down at the epoch of inflation, to within the uncertainties imposed by cosmic variance. In combination with new generations of deep wide field galaxy surveys, it should be possible to unambiguously measure the shape of the predicted peak in the power spectrum, and thereby establish unique constraints on the origin of the large-scale structure of the universe.