Fair enough. Though my point was trying to say that protocols should assume a variety of component failure modes.

I'd rather just have to trust the safety standards of the manufacturer of my car (and to a lesser extent the cars I might directly collide with), not the safety standards of every vehicle within transmit distance.