Originally Posted by CargoOne
Although I'm not a specialist in rocket science (or aircraft design) but generally speaking if certification authorities can accept the computer-generated changes into wing design as the proof of compliance to the standarts without any further actual tests why those tests have a place at all? Why not to accept the whole construction as "safe" based just on these computer models?
I am likewise neither of the above, but think I can offer at least part of the answer: The whole construction is a very complex system and the real-world/destructive testing may elicit flaws resulting from interaction of various parts of the whole, that the engineers did not anticipate. If destructive testing reveals a flaw that can be clearly identified as "simple," and confidently judged to have no interaction with other parts of the whole, and reliably reinforced to meet requirements by beefing up a single component or a few, then it seems to me that computerized or slide-rule engineering could supply a reliable result. If contrariwise the flaw cannot be identified as "simple," then it's back to the drawing boards. The more large-scale and complex the fix, the stronger the reason for a full-scale retest.