Making software more dependable
Software is integral in people’s daily lives. It supports everything from medical and tax document preparation, financial management, and web browsing on personal computers to controlling home appliance, medical device, and even vehicle operation.
It isn’t surprising, then, that software defects cause an enormous economic loss, currently in the hundreds of billions annually. There is therefore an urgent need for novel, effective techniques to ensure software systems are reliable, secure, safe, and maintainable—in short, that they are dependable.
Dr. Lin Tan, Canada Research Chair in Software Dependability, develops techniques and tools to study, predict, detect and fix software defects to improve software dependability.
Her research uniquely leverages interdisciplinary techniques including machine learning, natural language processing, data mining and program analysis to resolve software defects.
In addition, Tan’s work uses a new source of information—software text—to detect and fix software defects. Software text includes code comments, software documentation, and user manuals that are written in a natural language. By combining software text and traditional sources of information, such as source code, her work will achieve new breakthroughs in improving software dependability.
The timeliness and broad applicability of Tan’s research to the global software community are invaluable. Her technologies have already been transferred to multinational companies, and some have been integrated into leading software products.
Her research will continue to accelerate production of dependable code and directly help improve software to benefit millions of users across the globe.