Tracking the 'Life Cycle Trajectory' : Metrics and Measures for Controlling Productivity of Computer Aided Software Engineering (Case) Development

This paper proposes a new vision for the measurement andmanagement of development productivity related to computer aidedsoftware engineering (CASE) technology. We propose thatproductivity be monitored and controlled in each phase ofsoftware development life cycle, a measurement approach we havetermed life cycle trajectory measurement. Recent advances inCASE technology that make low cost automated measurement possiblehave made it feasible to collect life cycle trajectory measures.We suggest that current approaches for productivity managementinvolve the use of static metrics that are available only at thebeginning and end of the project. Yet the depth of the insightsneeded to make proactive adjustments in the software developmentprocess requires monitoring the range of activities across theentire software development life cycle. This can only beaccomplished with metrics that can measure performance parametersin each phase of the life cycle. We develop metrics that havethe ability to measure and estimate software outputs from eachintermediate phase of the development life cycle. These metricsare based on a count of the objects and modules that are used asbuilding blocks for application development in repository object-basedCASE environments. The viability of such object-basedmetrics for life cycle trajectory measurement has beenempirically tested for the software construction phase usingproject data generated in Integrated CASE developmentenvironments