Which inputs should a comparative test suite for transformed code prioritise?
Select an answer to reveal the explanation.
Short Explanation
You need representative production cases plus boundary and exceptional values. That's where transformed code tends to drift, especially around limits, rounding, and bad data.
Full Explanation
The correct approach is to prioritize representative production cases together with boundary and exceptional values, since those are where transformation differences typically appear. Typical cases confirm the main path, while boundaries such as field limits, zero and negative values, rounding points, empty records, and malformed records expose representation and default-behaviour differences. The choice that limits inputs to only the inputs used during the original program's development is incorrect because historical development inputs rarely reflect how the program is used now. The choice that uses randomly generated values only is incorrect because random values are unlikely to hit meaningful boundaries. The choice that uses the smallest possible input set is incorrect because optimizing for speed at the cost of coverage defeats the suite's purpose. For the exam, combine representative and boundary cases.