A study of productivity and efficiency for object-oriented methods and languages

A study was commissioned by the Hughes Space and Communications Software Engineering Group to determine the effectiveness of the recent introduction of object-oriented languages, technologies and development methodologies. Of particular interest was any effect on development productivity. Fundamentally, productivity metrics are difficult to apply across non-homogeneous projects and development teams. Furthermore, owing to many uncontrolled variables, such as the lack of a solid control project and non-rigorously collected data, productivity measures alone are were sufficient to determine meaningful results as requested for the study. A new, robust means of comparing non-homogenous development efforts (that does not require a control project) called “efficiency” was introduced and was used to augment the comparative analysis of the projects and address possible concerns with the use of productivity metrics alone. Efficiency measures the actual effort compared to an estimate of that effort with respect to an independent and well-defined baseline - for purposes of the study, this was COCOMO II. Conclusive evidence was found to support the hypothesis that object-oriented languages coupled with object-oriented methods result in greater productivity and efficiency as compared to other efforts. Furthermore, it is concluded that efficiency metrics, along with the non-standard use of COCOMO II, are a meaningful, useful and practical approach to compare development efforts