Case Based Reasoning System for Software Architecture Evaluation and Insight (CBR Insight)

ABSTRACT: The architectural complexity of a software code base has a significant impact on its operability, maintainability, and extensibility. In general, as the complexity increases, so do the number of bugs, the amount of time required to fix each bug, and the amount of time required to adjust existing features or add new features to the software. We propose to build the Case Based Reasoning System for Software Architecture Evaluation and Insight (CBR Insight), to provide contracting officials with a tool they can use to measure and understand the ramifications of architectural complexity.CBR Insight will combine the functionality of a leading COTS static analysis tool with case based reasoning techniques from artificial intelligence and leverage existing visualization and automation software. The result is a tool capable of generating the propagation cost and core size metrics and fully leveraging these results to provide insight into operational performance, risks, and costs.During Phase I we will develop use cases and requirements, design the CBR Insight system, build a limited prototype, evaluate the prototype on open-source DOD projects, gather feedback, and develop a plan for Phase II.; BENEFIT: The most direct targets for marketing the results of this effort are contracting officials and others involved in government and commercial software acquisition. CBR Insight would be especially attractive to this market given that it is an analysis tool designed for non-developers. By demonstrating significant results in evaluating and understanding the ramifications of software architectural complexity in Phase II, CBR Insight will provide the US Air Force with a strong incentive to fund the Phase III deployment of the proposed system.A secondary target of this effort is to market CBR Insight as an extension to the existing COTS code analysis tool that CBR Insight will be integrated with. CBR Insight would provide powerful new capabilities that would be of interest to current and new software developers utilizing their system. Additionally, any development team being evaluated on these architectural complexity metrics would need to put tools in place to monitor metric scores throughout the development process.