The very nature of the art object has resulted in slow progress toward electronic data processing in art museums. Art objects are described on several levels. The empirical, concerned with the piece as an object, and the main concern of the registrar, is most suitable for computer adaptation. This idea, however, is not an attractive one because the present manual system is seen as sufficient for collection manageability. The analytic and valuative levels are concerned with the object's meaning and significance. These are the curator's territory, and, since they are almost totally unsuitable for any computer adaptation, the curator finds computerization totally unsatisfactory for research and collection management. Some collections have tried computerizing records, but only when collection manageability, cost and effort-effectiveness have been facilitated to a greater extent by computerized systems will they be accepted. To date progress has been slow, but eventually collection size will force the issue of computerization. Only then will art museum networking, which is desirable in theory and follows computerization, become a reality. (Author/MBR)