The challenge of building process-control software

Process control is explained, and it is shown why trying to build process-control software using the same approaches developed and used for data-processing and information-system software will lead to failure. Rather than offering solutions to the problems inherent in developing process-control software, the author identifies the aspects of the problem on which to focus attention and the tasks that are ripe for solution. In particular, modeling and analysis, human-computer interaction, and risk evaluation are discussed.<<ETX>>