Four test-data categories
Each one catches a different bug class.
Cambridge expects you to use FOUR types of test data — each one exposes a different kind of bug.
Normal data. Valid data the algorithm should ACCEPT and process correctly. Confirms the everyday case works. Example for an exam-mark algorithm (0-100): 67.
Boundary data. Values right at the EDGE of the valid range — both just inside and just outside. Catches off-by-one bugs.
- Just INSIDE: 0 (lowest valid), 100 (highest valid).
- Just OUTSIDE: -1 (just below valid), 101 (just above valid).
Abnormal / erroneous data. Data that should be REJECTED — wrong type, completely out of range, wrong format. Confirms validation works. Example: 'A' (not a number), -5, 200.
Extreme data. Largest and smallest VALID inputs — used to stress-test. For our 0-100 mark example, the extremes coincide with the boundary inside-values (0 and 100). For a 'first-name' field, extremes would be 'A' (a single letter, valid) and a very long valid name.
Cambridge tip. Mark schemes consistently expect ALL FOUR types named with examples. Listing only normal data is a 1-out-of-8 answer.
- Normal — valid, processed.
- Boundary — at edges, both sides.
- Abnormal — rejected.
- Extreme — largest/smallest valid.