Sources of Errors & Solutions

Where things go wrong in a computer system — from a typo to a corrupted transmission — and the techniques used to catch and prevent those errors.

Sources of errors:

Category	Examples
Human error	GIGO (Garbage In, Garbage Out) — incorrect input leads to incorrect output; simple input mistakes (typos, wrong data entered)
Arithmetic errors	Rounding errors (small inaccuracies from rounding numbers), truncating (cutting off digits instead of rounding), fixed number of bits (a value too large for the allocated bits causes an overflow error — see 11.4.2)
Data transmission errors	Caused by physical issues like damaged/faulty cables disrupting the signal
Programming errors	Undetected logical errors — the program runs without crashing but produces incorrect results because the logic itself is flawed

Solutions for errors — verification vs validation is a critical distinction:

Verification: Confirming that data was entered/copied correctly, e.g. asking a user to type their password twice and checking the two entries match.
Validation: Checking that entered data is reasonable/acceptable according to defined rules, e.g. checking an age is between 0 and 120. Validation does NOT guarantee the data is actually correct/true — just that it's plausible.

Techniques for input that reduce human error at the source:

Barcode scanner / QR codes — scan pre-encoded data instead of typing it manually.
GUI design — using drop-down boxes, check boxes, radio buttons instead of free-text fields to limit input to valid options.
RFID (Radio Frequency Identification) — reads data wirelessly from a tag, e.g. access cards.
Biometric input — fingerprint/facial recognition for accurate identification.
OCR (Optical Character Recognition) — converts scanned/photographed text into editable digital text.

Types of data validation checks (also see 11.4.8, 11.4.14):

Check	What it verifies
Presence check	A required field has not been left empty
Range check	A value falls within an acceptable minimum and maximum
Uniqueness check	The value doesn't already exist elsewhere (e.g. no duplicate ID numbers)
Length check	The data has an acceptable number of characters
Type check	The data is of the expected data type (e.g. a number field doesn't contain letters)
Logical check	The value makes logical sense in context, e.g. an end date must be after a start date
Check digit	An extra digit calculated from the other digits, used to detect input errors, e.g. in barcodes and ID numbers
Checksum	A calculated value used to verify that a larger block of data hasn't been altered/corrupted

Data transmission check — parity: an extra bit added to a group of bits so the total number of 1-bits is always even (or always odd); if the parity doesn't match on arrival, an error is detected during transmission.

💡 Exam Tip

Verification and validation are constantly confused by students — remember: Verification checks data was captured/copied correctly; Validation checks data is reasonable according to rules. Neither guarantees the data is 100% true.