Technical Risk EstimationThe technical risk estimation project provides collaborative support for project managers and developers tasked with evaluating scope and impact o proposed changes as part of “What-If?” scenario planning, in terms of the probability and cost of cascading changes.
In this project, we are attempting to provide a technique and tool to allow early estimation of technical risk—the product of the probability that a change will be need and the cost of that change—that can be used for vertical communication within an organization.
The TRE tool represents a system as a graph where the nodes are classes and the arcs are dependencies. A user can select one or more points as the source of changes. TRE determines how those changes are likely to propagate, and colours the graph (in shades of red) to show the points that re most likely to be affected.
The basic concept has been published. We are also pursuing a variety of refinements to our basic technique and tool. Some of the work-in-progress: Support for systems written in multiple languages; improving historical data collection by accounting for past refactorings.
Robert J. Walker, Reid Holmes, Ian Hedgeland, Puneet Kapur, and Andrew Smith. A lightweight approach to technical risk estimation via probabilistic impact analysis. In Proceedings of the Mining Software Repositories Workshop at the International Conference on Software Engineering, 2006.
Robert J. Walker, Reid Holmes, Ian Hedgeland, Puneet Kapur, and Andrew Smith. A lightweight approach to technical risk estimation via probabilistic impact analysis. Technical report 2006-817-10, Department of Computer Science, University of Calgary, February 2006.