This paper was originally presented at an organizational meeting of a joint RAND-Stifftung Wissenschaft und Politik project on "The History of the Strategic Arms Competition," held in Washington, D.C., on November 15, 1985. It considers the defense planning process, both historically, and as a basis for making future decisions. The author contends that difficulties with nuclear planning cannot be attributed to the routine flaws that bedevil all governmental activities in a democracy, nor to the inability to comprehend the nature of modern weapons or to assess the risks posed by the Soviet nuclear establishment. He suggests that the key to improved nuclear planning lies in recognizing that nuclear weapons represent an absolute difference in kind from other instruments of force, and in developing a planning system that can function independently of our statements of national objectives. We need to develop surrogate "objectives" and new criteria for measuring military outcomes in the modern planning context.