Constructing a productivity model for small enhancements

Abstract

Software is in constant evolution and many approaches have been suggested to study software maintenance productivity. This master thesis aims to describe the process to design and implement a productivity measurement model for small functional enhancements to legacy software. Two motivations in uence this research: (i) understanding the productivity of the software maintenance process to help manage the cost of maintenance, (ii) understanding the cost drivers that a ect the software maintenance productivity. Another purpose of this research is to determine whether it is possible to construct productivity model(s) for small enhancements that is as good as development model. The case study, to show the feasibility of the application of the methodology, is coming from a large technology group that works for a bank organization. In this approach, it is showed that small enhancements can be measured by using the COSMIC measurement method that has the possibility to capture a lower level of granularity. Also, the quality of documentation for each small enhancement was studied. Because the person who has done the maintenance was at hand for this exercise, it was possible to complete the documentation. Therefore the quality of the documentation is high. Totally, 88 small enhancements are investigated and measured separately. The measurement of the functional size of each individual enhancement was based on the COSMIC ISO 19761 international standard. It was possible to produce a sound productivity model with this sample using all independent variables (R square of 0.75). The methodology used is showing that the experimentation was well controlled: within an enterprise for one major application for a period of time, design, program and implement by the same person, documented from the maintainer, measure within a control environment and verify by an expert.