Specification Mining with Few False Positives

Formal specifications can help with program testing, optimization, refactoring, documentation, and, most importantly, debugging and repair. Unfortunately, formal specifications are difficult to write manually and techniques that infer specifications automatically suffer from 90–99% false positive rates. Consequently, neither option is currently practiced for most software development projects. We present a novel technique that automatically infers partial correctness specifications with a very low false positive rate. We claim that existing specification miners yield false positives because they assign equal weight to all aspects of program behavior. For example, we grant less credence to duplicate code, infrequently-tested code, and code that has been changed often or recently. By using additional information from the software engineering process, we are able to dramatically reduce this rate. We evaluate our technique in two ways: as a preprocessing step for an existing specification miner and as part of a novel specification inference algorithm. Our technique identifies which traces are most indicative of program behavior, which allows off-the-shelf mining techniques to learn the same number of specifications using 60% of their original input. This results in many fewer false positives as compared to state of the art techniques, while still finding useful specifications on over 800,000 lines of