Murine, Gerald E.; Murine, Brian J. (1997, American Society for Quality) METRIQS International; San Juan Capistrano, CA; Boeing, Huntsville, AL
QM3 (Quality Metric Maturity Model) supports levels or sets of conditions for improving software quality metric modeling and implementation. This model consists of scenarios that generate quality metric data sets, such that the lowest level is the least mature in terms of impact, improvement, and usefulness. A continuous improvement model for software quality metric modeling and implementation includes steps such as: selecting quality attributes and a modeling framework; conducting a software quality metric training course; choosing a life cycle phase; collecting data; analyzing and reporting results; and implementing the results. If the improvement model is iterative, the next step is to choose the subsequent life cycle phase and then continue data collection and analysis. The QM3 is linked to life cycle phases or benchmarks. There are four QM3 levels, in which each succeeding level adds information to the earlier levels. Level 1 is a self-contained, non-calibration level. It requires collection of a minimum of data, but such collection should be complete. Level 2 adds data collected from software and engineering trouble reports. Level 3 incorporates productivity or cost-per-return data, such as information on cost, effort, re-work, size, and testing time. Level 4, the full maturity level, gathers trouble report and other independent data into one collection, such that the two sub-levels in levels 1, 2, and 3 do not exist in level 4. The large collection of information gathered at level 4 is a key to continuous improvement.
Data analysis,Data collection,Software quality,Process improvement,Product development,Metrics