CAIE 9618 · Paper 2 · Topic 2.5.1

Testing
Strategies

Black Box · White Box · Test Data Types · Test Tables · Trace Tables

CSZone Cambridge International AS & A Level Computer Science 9618

Why Test Software?

Finding Errors Before Users Do

Testing is the systematic process of running a program with planned inputs to verify it produces the correct outputs and handles all cases properly. The goal is to find and fix bugs before deployment.

TYPES OF ERROR

Syntax error — invalid code structure (caught at compile/parse time).
Runtime error — crash during execution (e.g. division by zero).
Logic error — program runs but produces wrong output.

TESTING APPROACHES

Black box — test without knowledge of internal code. Focus on inputs and expected outputs.
White box — test with knowledge of code. Ensure all paths/branches are covered.

Test Data Categories

Normal, Boundary, Erroneous & Extreme

For a function that accepts an age 0–120:

Type	Definition	Example (age 0–120)	Expected Result
Normal (valid)	Typical, expected input within range	25, 67, 5	Accepted, processed normally
Boundary	Exact at the edge of valid range	0, 120	Accepted (boundary included)
Extreme	Valid but at extreme ends of range	1, 119	Accepted and processed
Erroneous (invalid)	Outside valid range, wrong type, no input	-1, 121, "abc", empty	Rejected with error message

Key: Always test both sides of each boundary — just inside valid range AND just outside.

Test Tables

Documenting Tests Systematically

Test program accepting age 0–120 and outputting "Valid" or "Invalid":

Test #	Type	Input	Expected Output	Actual Output	Pass/Fail
1	Normal	45	"Valid"	"Valid"	✓ Pass
2	Boundary	0	"Valid"	"Valid"	✓ Pass
3	Boundary	120	"Valid"	"Valid"	✓ Pass
4	Erroneous	-1	"Invalid"	"Invalid"	✓ Pass
5	Erroneous	121	"Invalid"	"Valid"	✗ Fail
6	Erroneous	"abc"	Error message	Crash	✗ Fail

Exam Practice

Cambridge-style questions

Question 1

A program accepts a score from 1 to 100. State one example of each of the following test data types and explain why you chose it: (i) boundary data, (ii) erroneous data.

4 marks

2 marks

(i) Boundary data: 1 or 100 — because these are at the exact limits of the valid range; testing boundary values checks that the comparison operators use <= and >= correctly rather than < and >.

2 marks

(ii) Erroneous data: 0 or 101 or "hello" — because these are outside the valid range or of the wrong type; testing erroneous data confirms the program correctly rejects invalid inputs and displays an appropriate error message rather than processing them.

Common Mistakes

Don't lose easy marks

Giving only one boundary value — for a range 1–100, both 1 AND 100 are boundary values. Also consider 0 and 101 (just outside boundary). The mark scheme often expects candidates to identify the boundaries on both sides.

Confusing black box and white box testing — black box: tester doesn't see the code, tests from external specification (what the program should do). White box: tester sees the code, ensures all branches/paths are executed. A common mark-scheme requirement is to explain what the tester does or doesn't know.

Leaving the Actual Output column blank — in a test table question, the exam often presents a partly-filled table and asks you to complete it. If actual output is the same as expected, write it in — don't leave it blank. A blank actual output loses a mark.

Topic Summary — 2.5.1

What You Need to Know

TESTING TYPES

Black box: input/output, no code knowledge
White box: code coverage, all paths tested
Unit: individual modules · Integration: combined

ERROR TYPES

Syntax: invalid code structure
Runtime: crash during execution
Logic: wrong output, program runs

TEST DATA TYPES

Normal: typical valid input
Boundary: exact edge of valid range
Extreme: valid but at limits
Erroneous: outside range / wrong type
Test table: input, expected, actual, pass/fail

CSZone

Next Video

2.5.2

Program Design Methodologies

Waterfall · Agile · RAD · Software Life Cycle

Head to CSZone.co.uk for the complete worksheet, quiz, and interactive tools