Automated Explanations for Debugging

点击翻译按钮获取中文摘要

In this proposal, we focus upon the problem of automated debugging. The goal in this area is to devise techniques that help software developers understand and correct flaws within misbehaving computer programs. Existing techniques for automated debugging focus on making hypotheses about which lines of source code may contain a bug amongst all source code for a program. This proposal instead focuses on clearly and concisely explaining how and why a program misbehaved. These explanations are produced without human interaction and help to simplify and remove guesswork from the process of fixing faulty software. It is not presently known how to compute such explanations for programs involving concurrency, such as those with multiple concurrently executing threads, yet concurrency is emerging as a fundamental construct in modern programs and is of critical importance. This proposal addresses the heretofore unresolved issue of producing such explanations for programs involving multiple concurrently executing components. This proposal further addresses the important issue of presenting explanations to software developers in languages and formats that developers can understand and use. Improving the reliability and cost-effectiveness of software development has been a primary focus over the last decade for key software companies like Google, Apple, Microsoft, and IBM. Improved debugging tools lower the costs and increase productivity within software development. These lower costs and increased productivity are important for maintaining Canada's growing software industry. These techniques also improve the reliability of software, providing better assurances that computer programs do what they are intended to. As a result, companies and people who use the software also save money by avoiding hassles created by software that misbehaves.