ABSTRACT
The fitting of data to a regression line is now almost invariably carried out by the use of a computer. This essentially consists of an arithmetic/logic unit that manipulates data held in memory. Instructions control where the data are held, what is done with them, and where the result is to be stored. Such instructions are termed algorithms, which can be defined as a series of steps that achieve a desired aim. Computers work because of hierarchies of and interactions between algorithms, all of which must be devised by human beings. The range of tasks which can be carried out by a conventional computer is limited to those for which algorithms can be devised, and every step of the process must be spelt out to it.
