The problem of concurrency control in distributed databases is very complex. As a result, a great number of control algorithms have been proposed in recent years. This research is aimed at the development of a viable categorization scheme for these various algorithms. The scheme is based on the theoretical concept of serializability, but is qualitative in nature. An important class of serializable execution sequences, conflict-preserving-serializable, leads to the identification of fundamental attributes common to all algorithms included in this study. These attributes serve as the underlying philosophy for the categorization scheme. Combined with the two logical approaches of prevention and correction of nonserializability, the result is a flexible and extensive categorization scheme which accounts for all algorithms studied and suggests the possibility of new algorithms.