This paper focuses on the question of what needs to be predicted and what processes need to be understood to predict and forecast space weather conditions that are hazardous to current technology. The paper's aim is to see if we are working on the correct space parameters to permit prediction of those quantities that actually present hazards to current space technology. The paper is not intended to be encyclopedic. We conclude that although the sunspot number is a general proxy for many space hazards, there is surprisingly little direct need for its accurate prediction or for the prediction of solar flares as such. We also find that knowledge of Kp and other geomagnetic indices are rarely directly required. Important gaps in our knowledge exist concerning the variations of storm time electron, proton, and ion populations within the magnetosphere. Work is also required in predicting fast coronal mass ejections (CMEs) and in understanding the processes of CME initiation and acceleration within the corona and high energy solar particle acceleration and propagation. ¿ 2000 American Geophysical Union