To understand the ocean, its dynamics, its role in climate, weather and other ocean-atmosphere phenomena, one must observe it on a basin-wide scale with adequate time and space resolution. No such observation system yet exists. In the past few years, a number of technical developments have taken place that might permit the establishment by the 1990s of an observational system approaching these requirements. We focus here on two major components of such a hypothetical system: ocean acoustic tomography, and satellite observations of sea surface topography and of wind stress. These, combined with other types of observations and with a sensibly designed modelling effort, might provide an ocean-observing system at not unreasonable cost. This would lead to significant progress in ocean physics and dynamics, and one could contemplate realtime forecasting for ship routing, military purposes, fisheries, weather, and climate.