Which task should an engineer complete before an optimization suggestion for a mainframe routine is accepted into a production rollout?
Select an answer to reveal the explanation.
Short Explanation
You need to prove the optimized routine still produces the same results, not just that it runs faster. Same output first, better performance second.
Full Explanation
The correct task is to confirm through testing that the optimized routine still produces the same results as before, in addition to running faster. The task before accepting any optimization is to test that the routine's output has not changed, since performance is being improved on top of behaviour that must stay intact.
Confirming that the suggestion came from the analysis tool's highest-confidence category and applying it on that basis alone is incorrect because a confidence label from the tool describes the suggestion's likely quality, not whether it was verified against the actual codebase.
Removing any code the suggestion did not specifically flag, to simplify the routine further, is incorrect because deleting unflagged code goes beyond what the suggestion covers and introduces risk the optimization advice never addressed.
Confirming that the change reduces runtime by a fixed percentage before considering it acceptable is incorrect because chasing a fixed performance target says nothing about whether behaviour was preserved, and can encourage skipping validation to hit the number.
For the exam, the acceptance gate for any optimization is behavioural equivalence, confirmed by testing.