The introduction of desktop computing into primary healthcare over the past 30 years has been founded on the assumption that it would be accompanied by a clearly demonstrable benefit both to direct patient care, and to the development of administrative systems. While advances in primary care administration may be related to computerization, the evidence for improvements in direct patient care is more mixed. In this paper we critically discuss issues raised by the existing research literature in primary care computing, and especially note the difficulties posed by the absence of portable generic electronic patient records.