I do understand the whole generic Spy vs. Spy game (not the two birds in Mad mag.). My point is the in secure communications, the periodic generation of new keys is the norm. After 3 years, any enemy not bright enough to change their keys dozens of times, is no threat. They would have been eliminated early on.
"With regards to your last question, I believe Padilla is accused of plotting to detonate a radiological weapon in the city of Chicago."
Are you aware that the government has dropped (or attempted to) that charge?.
The cryptography analogy is not meant to be construed in a strictly literal fashion. It is not merely the breaking of encryption to which the methods of codebreaking apply; it is the entire range of intelligence-gathering activities, including placement of human agents, bugs, and other information-gathering tools. Whether we know what we know about Padilla via breaking an encrypted code or by a human agent in the enemy camp, or by mechanical spying devices, the important part is
not to let the enemy know what we know and how we know it. This is a fundamental principle in warfare.
I am aware of the status of the charges against Padilla. It is precisely because the government is making the same calculation I have been describing and has decided that the risk of revealing our intelligence methods is higher than the risk of Padilla as an individual going free, or being charged with less than his full crimes - and this is the correct decision to make, when faced with the difficult choice imposed on it by civil courts acting outside their authority.