THE PROPOSED PROJECT WILL DEVELOP A NEW TECHNIQUE FOR INVARIANT INFERENCE AND EMBED THIS AND OTHER CURRENT…

Public award description

THE PROPOSED PROJECT WILL DEVELOP A NEW TECHNIQUE FOR INVARIANT INFERENCE AND EMBED THIS AND OTHER CURRENT INVARIANT INFERENCE AND CHECKING TECHNIQUES IN AN EASY-TO-USE TOOL. THE RESULT WILL ENHANCE AN ENGINEER'S ABILITY TO USE FORMAL METHODS GENERATING, EDITING, REVIEWING, PROVING AND TESTING INVARIANTS AND IMPROVE PRODUCTIVITY IN VERIFICATION AND VALIDATION OF SAFETY AND CORRECTNESS PROPERTIES SOFTWARE. CURRENTLY, INVARIANTS THAT REPRESENT SUCH PROPERTIES REQUIRE EXTENSIVE HUMAN EFFORT TO WRITE; AUTOMATED TECHNIQUES, THOUGH IMPROVING, ARE STILL INSUFFICIENTLY CAPABLE OF AUTOMATICALLY INFERRING THEM. THE PROPOSED PROJECT WILL DEVELOP INNOVATIVE TECHNIQUES TO INFER LOGICAL INVARIANTS DESCRIBING THE BEHAVIOR OF INDIVIDUAL SOFTWARE MODULES BY COMBINING STATIC (ANALYZING A PROGRAM WITHOUT RUNNING IT) AND HYBRID ANALYSIS (INFERRING INVARIANTS FROM OBSERVATIONS OF EXECUTING SOFTWARE). IN PARTICULAR, THE PROJECT WILL (A) COMBINE CONCOLIC EXECUTION AND HYBRID ANALYSIS TO FIND CANDIDATE INVARIANTS FROM HIGH-BRANCH-COVERAGE TEST SUITES, (B) APPLY THAT COMBINATION TO OBTAIN INVARIANTS FOR INDIVIDUAL FUNCTIONS AND DATA STRUCTURES, (C) ITERATE THE ANALYSIS TO BROADEN DATA COVERAGE OF THE TEST SUITE AND IMPROVE THE ACCURACY OF INVARIANTS, AND (D) CREATE EARLY PROTOTYPES AND DEVELOPMENT PLANS TO INTEGRATE THE RESULTING TOOLS IN SELECTED IDES (ECLIPSE AND GRAMMATECH'S CODESONAR TOOL). IN CARRYING OUT THIS PROJECT, GRAMMATECH WILL BUILD ON ITS STATIC ANALYSIS TOOLS, CONCOLIC ENGINE, AND SOFTWARE DYNAMIC TRANSLATION MODULE. IT WILL LEVERAGE ITS BASE OF RESEARCH AND EXPERTISE IN STATIC AND HYBRID ANALYSIS, SPECIFICATION LANGUAGES, AUTOMATED SMT THEOREM PROVERS, AND GUI TOOLS FOR PROGRAM ANALYSIS AND DEVELOPMENT. THE COMMERCIALIZATION PROSPECTS FOR THE PROPOSED PROJECT ARE ENHANCED BY GRAMMATECH'S DEMONSTRATED EXPERIENCE IN PRODUCING PROTOTYPES AND COMMERCIAL PRODUCTS FROM RESEARCH RESULTS.