How might an ideal U.S. national security policy–making process function? For what reasons might the process deviate in reality from this ideal?