【第一場】
講 者：Prof. W. Eric Wong, Department of Computer Science, University of Texas at Dallas
講 題：Program Debugging with Effective Software Fault Localization
日 期：101年12月21日(五)
時 間：10:10~12:00
地 點：工程三館 115教室
主持人：謝續平教授
大綱：
Program debugging is an important activity in the development of reliable and trustworthy software. This activity can be very expensive and time consuming due to the difficulty of identifying the exact locations of program faults. In this talk, I will explain how to take advantage of sophisticated analysis of the static and dynamic behavior of the software under examination to prioritize suspicious code in terms of its likelihood of containing program bugs. Code with a higher risk can then be examined before that with a lower risk, as the former is more suspicious (more likely to contain program bugs) than the latter. Different techniques for fault localization based on execution slicing, code coverage, heuristics, neural networks, statistical analysis and fuzzy clustering will be discussed. Empirical data from case studies on programs with single and multiple bugs is used to show that our techniques outperform others that have the same goal.
【第二場】
講 者：Prof. W. Eric Wong, Department of Computer Science, University of Texas at Dallas
講 題：Challenges in Automatic Test Generation
日 期：101年12月21日(五)
時 間：13:20~15:10
地 點：工程三館 345教室
主持人：謝續平教授
大綱：
One of the most significant challenges in assuring the safe and secure operation of a software system is how to generate a small suite of test inputs designed to reveal program bugs, safety violations, or security vulnerabilities both efficiently and effectively. In this talk, we will discuss how to use sophisticated static and dynamic analyses to automate the test generation process for structural-based coverage testing (such as all nodes and all edges) and CEFSM-based communication testing (such as send-receiver pair, round-trip, and n-step message transfer). An innovative approach for testing software safety will also be discussed. The same approach can be extended to test software security.