We really started a process with the program review in the mid-1990s, and I've always argued that it has to be an ongoing process. Every time you create a new program, by definition, all of the previous programs have slipped down one degree of priority and they all have to be looked at. The new program should ultimately mean the elimination of an existing one. When a new program comes in, it has to be evaluated after two or three years to see if it's working. If it's not working, then get out of it.
On November 23rd, 2006. See this statement in context.