I've always wondered: is it the limitations of the sandbox that are the problem, or is it expectation? If we never had access to execute outside the sandbox, would we simply say we "see the reasoning" as we do for iOS?