Peter Gibbons Memorial Lecture Series: Making Software Testing Easier
Speaker: Myra B. Cohen, University of Nebraska
When: Refreshments at 5.30pm, lecture starts at 6.00pm.
Myra Cohen is an Assistant Professor in the Department of Computer Science and Engineering at the University of Nebraska at Lincoln. She received her PhD from the University of Auckland under the guidance of Professors Peter Gibbons and Rick Mugridge. Her research interests include software testing of highly configurable systems, software product line testing, and search-based software engineering. Her work has been applied in both academia and industry. She is the recipient of a National Science Foundation Early CAREER award and an Air Force Young Investigator award.
Synopsis: Software systems today are magnitudes of order larger and more complex than their recent ancestors. They are distributed, compositional and rely on inter-connectivity. Instead of building single systems, we now build families of systems and the software/hardware interfaces we once kept distinct have become blurred. As glitches in these large-scale systems continue to make newspaper headlines, developing reliable and affordable software presents an increasing number of challenges. The software engineering community is faced with a need for new and improved theories, techniques and tools to decompose this evolving web of complexity.
In this talk we examine challenges facing modern software systems in ensuring their dependability. We focus on the difficulty caused by one simple but ubiquitous concept; system configurability. Configurable systems include user configurable software such as web browsers, office applications that allow one to change their program preferences at will, families of products customized by businesses for different market segments, and systems that dynamically reconfigure themselves on the fly. The question that we are addressing is how can we efficiently test different configurations while avoiding unnecessary duplication of effort. In the spirit of practice embracing formalism, a theme that Peter Gibbons embodied, we show how architectural models, systematic sampling and guided search bring us closer to providing solutions for this difficult problem. The work presented in this talk began at the University of Auckland while the speaker was a student of Peter Gibbons.