The data processing problems which arise in the provision of a modern haematology service are discussed in the light of available resources, and the basic requirements of any automated data processing system are defined. The differences between the needs of haematology and those of other pathology specialities are emphasized. A strategy for the development of a simple system which covers data capture, analysis, and retrieval is described. Separation of the short-and long-term functions of the system provides a relatively simple, inexpensive solution which can be modulated to meet the particular circumstances of each laboratory.