There is a continuing growth of interest in the computer simulation of materials at the atomic scale, using a variety of academic and commercial computer programs. In all such programs there is some physical model of the inter-atomic forces, which may be based on something as simple as a pair interaction, such as the Lennard-Jones model, or as complex as a self-consistent, all-electron solution of the quantum mechanical problem. For a student or researcher, the basis of such models is often shrouded in mystery. It is usually unclear how well founded they are, since it is hard to find a discussion of the physical assumptions that have been made in their construction. The lack of clear understanding of the scope and limitations of a given model may lead to its innocent misuse, resulting either in unfair criticism of the model or in the dissemination of nonsensical results.