ABSTRACT
Since errors are inevitable in numerical calculations, it is essential to have some basic concepts about various types of errors that are encountered and ways of estimating them. The errors can be grouped into three categories, namely inherent or experimental, truncation and roundoff. The absolute error may bear any sign and is defined as the difference between the true and the approximate values. In some texts it is defined the other way, that is, the difference between the approximate and the true values. So long as the absolute value of the absolute error is used, this difference in definition does not affect anything. For a true value close to unity, the absolute and relative errors are nearly equal, when the latter is expressed as a fraction. Otherwise, one may differ from the other by a large margin.
