Gregory Cochran discusses The Wizard War, by R. V. Jones — partly because it provides some good examples of thin and thick problem-solving:
Reginald Jones (Ph.D. Oxford, 1934) was one of the first scientists to work for an intelligence service. He investigated German radio navigational systems and developed various methods of interfering with them, which often involved projecting the German’s next move in the electronic war. He was one of the developers of chaff, and also served as an expert consultant on the development of German rocketry — mainly the V-2.
Some of his successes were classically thin, as when he correctly analyzed the German two-beam navigation system (Knickebein). He realized that the area of overlap of two beams could be narrow, far narrower than suggested by the Rayleigh criterion.
During the early struggle with the Germans, the Battle of the Beams, he personally read all the relevant Enigma messages. They piled up on his desk, but he could almost always pull out the relevant message, since he remembered the date, which typewriter it had been typed on, and the kind of typewriter ribbon or carbon. When asked, he could usually pick out the message in question in seconds. This system was deliberate: Jones believed that the larger the field any one man could cover, the greater the chance of one brain connecting two facts — the classic approach to a thick problem, not that anyone seems to know that anymore.
All that information churning in his head produced results, enough so that his bureaucratic rivals concluded that he had some special unshared source of information. They made at least three attempts to infiltrate his Section to locate this great undisclosed source. An officer from Bletchley Park was offered on a part-time basis with that secret objective. After a month or so he was called back, and assured his superiors that there was no trace of anything other than what they already knew. When someone asked “Then how does Jones do it?” he replied “Well, I suppose, Sir, he thinks!”