Auto-consistent environment for measurement software development

In a measurement environment, the user will repeatedly change some of the measurement settings and will want to view the results. Developers of elaborate measurement systems are often puzzled as to whether the generated results correspond to the latest settings. Data consistency in a measurement system means that displayed results will always conform to the latest settings. This paper suggests that automatic data consistency is an underestimated feature of dataflow environments. If only a few of the settings are modified each time, it is unreasonable to repeat operations that do not depend on the changed settings. A demand-driven execution scheme is presented in which data tokens are tagged with their time of calculation. This scheme avoids redundant operations without programmer intervention. This is an important concern when some real-time constraints are to be met, because the operations in complex measurement systems can be very time-consuming. When the developer must handle these two issues programmatically, their combination often results in unmanageable programs