Excellent computer simulations are done for a purpose. The most valid purposes are to explore uncharted territory, to resolve a well-posed scientific or technical question, or to make a design choice. Stand-alone modeling can serve the first purpose. The other two goals need a full integration of the modeling effort into a scientific or engineering program. Some excellent work, much of it related to the Department of Energy Laboratories, is reviewed. Some less happy stories are recounted. In the past, some of the most impressive work has involved complexity and chaos. Prediction in a complex world requires a first principles understanding based upon the intersection of theory, experiment and simulation.

30 pages, 8 figures