The purpose of this conference was to investigate a new and promising field in computer science: the application of quantitative, and particularly statistical, methods to the study of computer performance. The present state of the art is that on one hand is a wealth of data, if gathered in a rather haphazard fashion; on the other hand are theoretical models whose stringent assumptions are not examined and whose conclusions are not verified. The papers accepted attempt to deal with real data in a reasonably sophisticated manner. The organizers, speakers, and discussants hope to open up the field, enhance understanding of computer systems, and contribute to statistical methodology.