SHF: Small: User-Centered Software Analysis Tools

University Of Maryland, College Park, College Park MD

Investigators

Abstract

The proposed work explores techniques for making static analysis tools more effective in the hands of users. This will lead to the development of new static analysis algorithms, techniques, and interfaces that provide the information users need to find, verify, and fix defects. The core of the proposed work will be a framework for static analysis visualization and interaction, with several components. First, the framework will include checklists to help users triage defect reports, i.e., decide whether they are true or false positives. The aim is to develop ways to instrument static analyses to automatically generate checklists based on imprecision introduced during the analysis. Second, the framework will include lightweight query and search facilities to help users work with static analysis tool results. Users need effective ways to query the knowledge-base generated by a tool when trying to understand an error report; current static error reports tend to provide too little or too much information. Third, the framework will include a generic visualization for program paths, a core part of many static analysis tools' defect reports. While some tools include simple path visualization, our framework will aim for far more effective interfaces by applying information visualization principles.