FrançaisGet started

Test data sets

Designing a program's test data set: a test is an input and an expected result, with nominal cases, boundary cases and error cases, values chosen to reveal defects and a table of tests.

  • 25 cards
  • 5 topics
  • Free
Study this deck

The outline, topic by topic

Each topic is weighted the way the exam weighs it.

  1. Concepts
  2. Questions
  3. Sort and link
  4. Scenarios
  5. Understand

A few cards from the deck

  • Test case: what is it?

    Show the answer

    A precise input given to the program, paired with the result it must produce for that input.

    valeur_absolue(-3) must give 3: the test is the pair (-3, 3). A test data set is a list of such pairs, laid out in a table of tests.

  • Which test shows that a version which only checks a % 4 == 0 is wrong?

    ```python
    def est_bissextile(a):
    return a % 4 == 0 and a % 100 != 0 or a % 400 == 0
    `

    • 1900, expected False
    • 2000, expected True
    • 2023, expected False
    • 2024, expected True
    Show the answer

    1900, expected False

    1900 is divisible by 4 but not a leap year, since it is divisible by 100 without being divisible by 400: the simplified version would return True. The other three years give the right result with or without the century rule.

  • A test pairs an … with an …, set before the run.

    Show the answer

    A test pairs an input with an expected result, set before the run.

    Setting the result before running avoids being swayed by the program's output, whether right or not.

  • Expected result: what is it?

    Show the answer

    The value the program must return for a given input, set before running it from the specification.

    It is computed by hand or taken from the specification. Copying it from the program's output would prove nothing: the defect would be copied with it.

  • You are writing a test's expected result. Where should it come from?

    • From the specification or a calculation by hand, before running
    • From the program's output, copied after running it
    • From the result of another function of the same program
    • From the value that seems most likely after a glance at the code
    Show the answer

    From the specification or a calculation by hand, before running

    An expected value copied from the output is always True: if the program is wrong, so is the test. The expected value measures the program because it comes from somewhere other than the program.

Study this deck