Experience Report: Using RESOLVE/C++ for Commercial Software

摘要

Academic research sometimes suffers from the "ivory tower" problem: ideas that sound good in theory do not necessarily work well in practice. An example of research that potentially could impact practice over the next few years is a novel set of component-based software engineering design principles, known as the RESOLVE discipline. This discipline has been taught to students for several years 23, and previous papers (e.g., 24) have reported on student-sized software projects constructed using it. Here, we report on a substantial commercial product family that was engineered using the same principles - an application that we designed, built, and continue to maintain for profit, not as part of a research project. We discuss the impact of adhering to a very prescriptive set of design principles and explain our experience with the resulting applications. Lessons learned should benefit others who might be considering adopting such a component-based software engineering discipline in the future.