Description of errors in data processing

1. Computational errors
- Occurs when an arithmetic operation does not produce the expected results. The most common computation errors include overflow, truncation and rounding
- Overflow errors
- Occurs if the result from a calculation is too large to be stored in the allocated memory space. For example if a byte is represented using 8 bits, an overflow will occur if the result of a calculation gives a 9-bit number.
- Truncation errors
- Result from having real numbers that have a long fractional part which cannot fit in the allocated memory space. The computer would truncate or cut off the extra characters from the fractional part. For example, a number like 0.784969 can be truncated to four digits to become 0.784
The accuracy of the computer output is critical. As the saying goes garbage in garbage out (GIGO),the accuracy of the data entered in the computer directly determines the accuracy of the information given out.

Some of the errors that influence the accuracy of data input and information output include

Transcription,
Computation and
Algorithm or logical errors.

2.Transcription errors

Occurs during data entry. Such errors include misreading and transposition errors
Misreading errors
Are brought about by the incorrect reading of the source by the user and hence entering wrong values. For example a user may misread a handwritten figure such as 589 and type S89 instead i.e. confusing 5 for S.

Transposition errors
Results form incorrect arrangement of characters i.e. putting characters in the wrong order. For example the user might enter 396 instead of 369.
These errors may be avoided by using modern capture devices such as bar code readers, digital cameras etc which enter data with the minimum user intervention.

Rounding errors
Results from raising or lowering a digit in a real number to the required rounded number. for example, to round off 30.666 to one decimal place we raise the first digit after the decimal point if its successor is more than or equal to five. In this case the successor is 6 therefore 30.666 rounded up to one decimal place is 30.7.if the successor is below five,e.g.30.635,we round down the number to 30.6

Algorithm or logical errors
- An algorithm is a set of procedural steps followed to solve a given problem. Algorithms are used as design tools when writing programs. Wrongly designed programs would result in a program that runs but gives erroneous output. Such errors that result from wrong algorithm design are referred to as algorithm or logical errors.