Speaker Alex Aiken
Affiliation Stanford University
Host John Spencer
Date recorded 17 January 2006
Despite years of research on static techniques, the state of practice for analyzing concurrent software remains relatively primitive. Recent advances in program analysis technology, however, hold out some hope for significant progress. This talk will present the approach and initial results of a project to build an effective static race detector, namely one that developers can use routinely to reliably identify dangerous races in realistic scale systems. We'll discuss how far we've advanced toward that goal as well as what problems remain. (Joint work with Mayur Naik.)
©2006 Microsoft Corporation. All rights reserved.