Darko Marinov

Ph.D. Massachusetts Institute of Technology, 2004

Research Statement

My main research interest is improving software reliability. My thesis research in automated testing focuses on structurally complex data that pervade modern software. Specifically, I developed techniques and tools for the efficient generation of bounded-exhaustive test suites from compact test descriptions. Deployment of these tools has enabled developers to find significant errors in several real applications. As another means for increasing reliability, I have used theorem proving in proof-generating compilation and for checking relational formulas. My research has also explored a variety of other areas, including profiling, specification languages, enumerative combinatorics, and computer architecture.

In the future, I plan to extend my work on software testing and develop new techniques that are easier to use and provide more efficient test generation. I intend to combine these techniques with model checking and constraint solving to target a wider range of applications. In the longer term, I intend to explore various dynamic and static analyses for checking properties of software. I believe that a fusion of dynamic and static approaches offers great opportunities for improving the reliability of production-scale software.