> 1. If the government claims "state secrets" they have a burden of proof to demonstrate a reasonable probability of specific and tangible harm actually occurri

> 1. If the government claims "state secrets" they have a burden of proof to demonstrate a reasonable probability of specific and tangible harm actually occurring. This is a terribly bad idea, if you consider the Curry-Howard-correspondance between proof and programm, and the constructive logic approach of proof by construction. That's a stretched metaphor, perhaps, but it's very real, considering that one often waits how things play out. Just like wikipedia let's everyone edit, see how it turns out, and revert later--which can be messy or impossible if several legit edits were made meanwhile.