|Williams Professor Wins Award for New Research in Computer Science|
|08:33PM / Thursday, August 29, 2019|
WILLIAMSTOWN, Mass. — Stephen Freund, professor of computer science at Williams College, received the 2019 Most Influential Paper Award at this year's Association for Computing Machinery (ACM) Conference on Programming Language Design and Implementation, the premier forum for researchers, developers, practitioners, and students to present research on computer programming languages.
This award recognizes research that has pushed forward the state-of-the-art, opened new research directions, and had a significant practical impact on the computing field as a whole over the past decade.
Freund and his co-author Cormac Flanagan (University of California, Santa Cruz) published the work leading to this award in 2009 in a paper titled "FastTrack: Efficient and Precise Dynamic Race Detection." The research in that paper developed a new technique for finding data race conditions, a particularly harmful type of computer bug.
"Race conditions occur when two threads running at the same time on a multi-core processor or multi-processor system manipulate a shared memory location without proper synchronization," Freund said. "The negative impacts of race conditions can range from data corruption to catastrophic system failure, and developing effective ways to detect when a race condition bug occurs has been an active area of research for several decades."
Freund's paper addresses the limitations of prior techniques to find race conditions, which have typically been too time-consuming to use or report too many false positives. False positives are problematic because they require programmers to invest time tracking down errors that do not actually exist.
"Our work on FastTrack changed that," Freund said. "We developed an algorithm that was efficient enough to use even on very large systems while still never under-reporting or over-reporting problems."
FastTrack was quickly and widely adopted within the computer science research community and industry, and the insights behind FastTrack have led to further advances on a number of other program-checking and verification problems.