Accurate measurements of productivity and quality are essential for balancing workload, creating predictable schedules and budgets, and controlling quality. Traditional software development processes include well-established methods for measuring productivity and quality. These include Lines of Code (LOC). A measure of the size of a software application, LOC is the foundation for productivity measurements (LOC/unit work) and quality measurements, such as defect densities (defects/LOC).
With the introduction of Model-Based Design, organizations require a different measure of the software development process. For example, Model-Based Design enables automatic code generation from graphical models. This means that the average engineer can produce remarkably more LOC per unit time than is possible with hand coding, with virtually no software coding defects. While these productivity gains are grounded in real process improvements, new metrics are required to properly instrument and measure those improvements.
The automatic capture of process metrics in a modern development process increases data accuracy and overall productivity. Practical experience has shown that an organization quickly learns to 'manage' metrics captured manually to produce mandated improvements, often without improving the underlying process. Additionally, requiring developers to manually capture a comprehensive set of process metrics can burden and distract them from their primary work. Model-Based Design offers the capability to automatically extract metrics, minimizing cost, time to market, and avoidance of quality-related issues.
The Modeling Metric tool provides an automatic, noninvasive measurement technique for gathering accurate metrics in the Simulink® and Stateflow® environment. |
