> (1) Churchill's Coventry quandary is a myth. It's not whether it actually happened that's important or not, and I think you realize that. Prime Minister Chur

> (1) Churchill's Coventry quandary is a myth. It's not whether it actually happened that's important or not, and I think you realize that. Prime Minister Churchill certainly did need to be careful about what actions he took based on ULTRA decrypts so as to avoid revealing the activities of Bletchley Park. Coventry is the quotable example despite not actually happening the way the story is told. However if you require an example that actually happened, consider when Adm. Nimitz received message intelligence pointing out Adm. Yamamoto's flight itinerary. In his case he ended up using a literal example of parallel construction (having a coastwatcher send a report of a Japanese flight) to generate a non-codebreaking reason for his staff to know about the flight, and then sent out his fighter planes to shoot Yamamoto's flight down. > (2) The NSA is way out of line in this respect and no matter how much they do/did it that stands and you should not apologize for it. With respect to things like DEA I would agree with you. But on the other hand if they came across a no-shit email from a hitman or something describing an upcoming murder I would expect them to report it. It may be that it's impossible to write a policy that straddles that border, in which case I'd say we should probably ban that type of information sharing and then rely on the idea that an analyst who sees that type of murder plot might still leave an anonymous tip. > The balance here is not to get to perfect security without privacy, the balance point is to get to an optimal level of security with a maximum of privacy. Hey look at that, we fully agree!