Purpose of Testing

For example, a program intended to add two numbers should produce 7 when given 3 and 4. If it displays -1, it has run, but its result is incorrect. Testing exposes the fault; the developer must then correct the code and test it again to check the correction.

This supports correctness: producing the intended results. Testing different inputs also helps assess reliability: whether the program consistently works as intended, rather than succeeding only in a particular case.

Handling unexpected input safely

Users do not always enter what a programmer expects. They may misunderstand an instruction, leave a field blank or type text where a number is needed. A robust program handles these situations without crashing or producing inappropriate results.

Consider a program that asks for a whole number from 1 to 10. It should accept 5, but reject "abc" safely, perhaps displaying a clear message and allowing another attempt. Testing should check both behaviours. Rejecting unsuitable input is a correct response, not a failure of the program.

Testing can also check security-related requirements. For example, a login system should refuse access when the password is incorrect. A successful login with the correct password alone does not show that this protection works.

Looking for faults, not just successes

Effective testing deliberately looks for situations in which the program might fail. This is sometimes called destructive testing: the aim is to uncover faults, not physically damage the computer or its data.

Choosing only inputs that already work can hide problems. A wider range of tests helps reveal incorrect outputs, crashes and unmet requirements so that developers can address them before users encounter them.

Testing increases confidence in a program, but passing the tests does not prove that every possible error has been removed.