Software reuse: data conversion experiences and issues

Abstract

At Boeing, the large, multiple-release Define and Control Airplane ConfigurationlManufacturing Resource Management (DCACYMRM) data conversion effort had to develop a complete conversion system that could: 1) load data extremely quickly, 2) handle large data volumes of legacy extracts, 3) adapt to architectural changes simply and easily, and 4) be used by programmers with no special training. The phased nature of the loads required reusable loading techniques, and the changing target architectures required adaptable techniques. In this paper, we discuss our approach to loading large volumes of data efficiently into a system during three distinct releases. As part of the solution, we implemented a repeatable software engineering process, in part because of the many shortcomings of using commercial-off-the-shelf (COTS) vendor’s approaches to creating the software. This paper will trace the adaptive capabilities of our software reuse techniques through the three releases. We highlight factors that allow the same reuse techniques to be used for repeatable processes in large industrial settings. Finally, we present observations about software reuse gathered from our experience. For Boeing to succeed, we had to employ a reuse approach to our software development. We settled on using adaptive frames to create accurate logic decisions in the code, and by optimizing the efficiency of the resulting code.