Making Sense of Requirements, Sooner
Tim Menzies, Julian Richardson · Computer · 2006
Making decisions during early requirements formulations is like racing through a dark forest without a road map or compass. The available options or goals are often unclear. Different participants in the formulation process have different points of view and frequently end up making decisions based on political rather than technical grounds. Using AI-based simulation tools can cut back the forest to reveal clear paths to possibilities and useful requirements. Researchers have long used modeling and simulation to analyze complex processes. In this approach, we describe the system's properties and behavior symbolically and numerically, and then let the computer unleash its brute force to grind through the combinations. We have been developing our own model-based technique for making early requirements decisions. The technique uses two curious properties seen in many models, clumps and collars, which make it easier to search quickly through a seemingly vast range of options