ABSTRACT
Computation, as its name suggests, involves numerical calculations. Numerical calculations can be in error due to ordinary arithmetical mistakes and/or due to the use of approximate values in the calculation. Arithmetical mistakes can be largely eliminated by a systematic approach to numerical work and by carrying out rough checks of the arithmetical results wherever possible. Approximate quantities are frequently used in computation. Scientific data such as length, mass, time, etc., can only be as accurate as the measuring equipment used and these quantities are therefore usually in error to some degree. Also, the results of arithmetic operations on some numbers can only be expressed as approximate values. The value of the error involved in using an approximation may be expressed in two ways: absolute error and relative error. The more complicated the calculation the more essential it becomes to predict the error range that the final answer must lie within.
