My point is there is no way to predict viability from code structure or "information".
If you want to make ID a science that would be a good line of research; demonstrate that you can make something entirely new that will survive. Demonstrate that you can predict the outcome of a previously untested mutation.
Well, our genetic programming abilities are still in their infancy, I admit.
Nonetheless, it is man's Intelligent Intervention, not natural, unaided "evolution" that is responsible for the examples that you mention above.