One approach to the metric baselining imperative for requirements processes

Abstract

The success of development projects in customer-oriented industries depends on reliable processes for the definition and maintenance of requirements. With the sustained, severe reduction in the rush to new technology, this widely accepted fact has become increasingly evident in the networking industry. Customers now focus on high product quality as they strive for economy of operation. Enhancing product quality necessitates enhancing processes, which in turn can necessitate applying more accurate (and precise) measures. Finding process deviations and identifying patterns of product deficiencies are critical steps to achieving high quality products. We describe the application of quantitative process control (QPC) during early development phases to establish and maintain baseline distributions characterizing RMCM&T processes, and to monitor their evolutions. Metric baselining as described includes key metric identification, and data normalization, filtering, and categorization. Empirical baselining provides the statistical sensitivity to detect requirements process problems, and to support targeted identification of particular requirements-related patterns in defects.