Validation, Verification & Testing (Years 9–11) — Computer Science

Checking data is sensible with validation, checking it was entered correctly with verification, and choosing normal, boundary and abnormal test data.

A program is only as trustworthy as the data you feed it. Type your age as "-7" or your email as "banana" and, without checks, the program will happily store nonsense — and rubbish in means rubbish out. Two guards stand at the door: validation, which asks "is this data sensible?", and verification, which asks "was this data entered correctly?". Then careful testing proves the guards actually work.

Why this matters: Validation and verification keep real systems safe and accurate, from online forms to bank records. Cambridge expects you to name validation checks, distinguish validation from verification, and choose normal, boundary and abnormal test data.

Validation asks if data is sensible; verification asks if it was entered correctly.

Validation checks

Validation is an automatic check the computer performs to make sure data is reasonable before it is accepted. It cannot prove data is correct — only that it is plausible.

Check	What it tests	Example
Range	Value between limits	Age 0 to 120
Length	Right number of characters	Password ≥ 8 characters
Type	Correct data type	Age must be an integer
Presence	A value was entered	Surname not left blank
Format	Matches a pattern	Date as DD/MM/YYYY
Check digit	Extra digit recomputes	ISBN, barcode

Verification

Double entry — the data is typed twice and the two copies are compared; if they differ, a typing slip has been caught (common for passwords).
Visual check — the user reads the entered data on screen and compares it against the original source document.

Note that verification does not make data correct either — you can carefully type the wrong number twice.

Test data: normal, boundary, abnormal

For a range 1–100: normal 50, boundary 1 and 100, abnormal 0 and 101.

Normal data — typical values that should be accepted (e.g. 50 for a 1–100 range).
Boundary data — values right on the edge of what is allowed (1 and 100 should be accepted; 0 and 101 should be rejected). Testing the boundary catches off-by-one errors.
Abnormal (erroneous) data — values that should be rejected (e.g. -5, 200, or the word "cat").

Testing a validation routine

A trace table lets you follow test data through a routine and check the output. Testing with all three types of data gives confidence the routine handles every case.

Test data	Type	Expected result
50	Normal	Accepted
1	Boundary (lowest valid)	Accepted
100	Boundary (highest valid)	Accepted
0	Abnormal	Rejected
150	Abnormal	Rejected

Worked example (State + apply, 3 marks): A field accepts marks from 0 to 100. Give one item of normal, boundary and abnormal test data.
Normal: 55 (a typical valid mark, accepted).
Boundary: 0 or 100 (the exact edges, accepted); or -1 / 101 (just outside, rejected).
Abnormal: 250 or "pass" (clearly invalid, rejected).

Common mistake: Saying validation guarantees the data is correct. It does not — "Jhon" passes a presence and type check but is still a misspelling. Validation checks data is sensible, not true. Verification (double entry) catches typing slips but still cannot prove truth.

Memory hook: Validation = Valid/sensible, Verification = Verify it was copied right. For test data remember N-B-A: Normal, Boundary, Abnormal.