Testing programs: test plans and test data

For an input that must be a whole number from 0 to 100:
Normal data: typical valid values that should be accepted, such as 55.
Boundary data: values at the very edge of the valid range, such as 0 and 100. Programs often fail here, for example by using < instead of <=.
Erroneous data: values that should be rejected, such as "abc", 150 or -5.

Each row has a test number, what is being tested, the test data, the expected result (worked out before running) and, after running, the actual result. If they differ, the program is fixed and tested again.
def grade(mark):
if mark >= 70:
return "Distinction"
elif mark >= 40:
return "Pass"
else:
return "Fail"

This loop rejects anything that is not all digits or is above 100:
mark = input("Mark: ")
while not mark.isdigit() or int(mark) > 100:
mark = input("Whole number 0 to 100: ")
mark = int(mark)
print("Accepted", mark)
With the inputs abc, 150, -5 and 72, it asks again three times, then prints: Accepted 72
"-5" is rejected because the minus sign is not a digit, so isdigit() is False.

Fit for purpose: it does everything the requirements ask, gives correct results for all the test data and handles bad input without crashing.
Efficient: it does not repeat work. For example, it stops a search once the item is found, uses a loop instead of copied lines, and does a calculation once after a loop rather than every time round it.
When asked to evaluate a program, say what works, what does not, and suggest a specific improvement.