The major types of cloud retrieval algorithms are reviewed with special emphasis being placed upon recent (i.e., post 1981/1982) developments and novel techniques. Satellite-based retrieval algorithms can be grouped into three classes: threshold methods, statistical procedures, and radiative transfer techniques, although each algorithm depends upon implicit, if not overt, inversion of the radiative transfer equation. The fourth type of retrieval differs very considerably from the satellite-based techniques as it depends upon surface-based measurements which are generally, but not always, human rather than computer based. There is a tendency to assume that surface-based observations, especially of total cloud amount, are "correct" but that they may differ from satellite-based retrievals because of the differences in viewing geometry. Actually the literature reveals surprisingly few intercomparison studies. None of the satellite-based techniques have yet been well-validated in a variety of situations, while surface-based observations are made in all terrain and climate regimes. Overall there seems to be good reason to believe that surface-based observations of cloud amount and especially of low cloud amount, character, and base height, can add significantly to satellite-based global nephanalyses now operational or currently being planned.