Improved software engineering decision support through automatic argument reduction tools
Tim Menzies, James D. Kiper, Martin S. Feather · 2003
Some of the most influential decisions about a software system are made in the early phases of the software development life cycle. Those decisions about requirements and design are generally made by teams of software engineers and domain experts who must weigh the complex interactions among requirements and the associated developmental and operational risks of those requirements. Some of these early life cycle decisions are more influential, or perhaps fateful, to subsequent software design and development than are others. When debating about complex systems with a large number of options, humans can often be slower than an AI system at identifying the clusters of key decisions that give the most leverage. By focusing a group of human domain experts or software engineers on these key decision clusters, more time can be devoted to these pivotal decisions and less time is wasted on irrelevancies. We are developing a tool based on an integration of: JPL’s DDP group decision support tool [3] WVU’s contrast set learners [7]) Miami University’s cluster visualization tools Various component of this tool have been tested on case studies at JPL. 1