Verification, Testing and Statistics

  • Authors:
  • Sriram K. Rajamani

  • Affiliations:
  • Microsoft Research, India

  • Venue:
  • ICTAC '09 Proceedings of the 6th International Colloquium on Theoretical Aspects of Computing
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Formal verification is the holy grail of software validation. Practical applications of verification run into two major challenges. The first challenge is in writing detailed specifications, and the second challenge is in scaling verification algorithms to large software. In this talk, we present possible approaches to address these problems: We propose using statistical techniques to raise the level of abstraction, and automate the tedium in writing detailed specifications. We present our experience with the Merlin project [4], where we have used probabilistic inference to infer specifications for secure information flow, and discovered several vulnerabilities in web applications. We propose combining testing with verification to help scalability, an reducing false errors. We present our experience with the Yogi project [1,2,3,5], where we have built a verifier that combines static analysis with testing to find bugs and verify properties of low-level systems code.