Interchanges between workers in different areas of computer science have become more frequent in recent years. In this paper, five people working in artificial intelligence attempt to relate their experience to the development of simulations. The paper is really five separate short papers, one by each of the authors, linked together by this common theme.

The first segment is written informally and from a broad perspective. Later segments are somewhat more formal and emphasize specific issues. The second segment emphasizes the importance of sensitivity analyses, and the third emphasizes the importance of a thorough knowledge of the problem domain and recommends using non-quantitative information. The fourth segment also recommends using non-quantitative information and emphasizes the importance of validation. The last segment recommends using non-quantitative and uncertain information.

The paper is relatively easy to read and contains some useful suggestions. Most of these suggestions, however, such as the use of sensitivity analyses, the importance of validation, and the use of uncertain information, have been a part of good simulations from the start. Non-quantitative information has not been widely used in simulations, but it is difficult to see exactly how this might be done, and none of the authors provides an example of its use in a simulation.