软件崩溃的分析，聚类和调试技术研究

项目名称： 软件崩溃的分析，聚类和调试技术研究

项目编号： No.61272089

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 张洪宇

作者单位： 清华大学

项目金额： 81万元

中文摘要： 随着软件产业的发展，软件系统正变得越来越复杂，软件失效的可能性也越来越大。在各样软件失效中，软件崩溃（Software Crash）造成的影响相对较大。软件崩溃是指软件在运行过程中不正常地停止工作的现象。现在工业界广泛采用的软件质量保证技术很难保证软件崩溃不会发生。同时，随着软件崩溃报告的数量的增加，传统的崩溃处理方法已经很难适应。在本项目中, 我们将研究有效的软件崩溃分析、聚类与调试方法。我们将深入分析崩溃产生的原因及特性，并提出相似崩溃报告的聚类方法（即将由相同的软件缺陷造成的崩溃报告聚成一类）。我们也将提出崩溃的重现方法（即帮助程序员快速复制、重现崩溃情形），和导致崩溃的缺陷的定位方法（即帮助程序员找到出错的地方）。我们将在实际软件项目中验证所提出的方法。我们相信关于软件崩溃的研究能够帮助企业提高软件质量，减少崩溃带来的损失。

中文关键词： 软件崩溃；软件测试；软件缺陷管理；软件数据挖据；软件维护

英文摘要： With the rapid development of software industry, software systems are getting more and more complex and error-prone. Although project teams spend much resources and effort on software testing before releasing products, in reality, released software still contains bugs. Some bugs manifest as crashes in the field. Software crash means unexpected termination of a software system during runtime. With the increasing number of crash reports, traditional crash-handling methods should be improved. In this project, we will explore methods for automated analysis, clustering and debugging of software crashes. We plan to analyze the possible causes and characteristics of software crashes, and a method for clustering similar crash reports that are caused by the same bug. We will also propose methods for reproducing crash reports (to help developers replicate the crashing scenarios), as well as methods for locating crashing faults. We will evaluate the proposed method on large-scale, real-world projects. We believe such a project could help industry improve software product quality and reduce the damage caused by software crashes.

英文关键词： Software Crash；Software Testing；Software Defect Management；Mining Software Repository；Software Maintenance