METAPHOR BASED LANGUAGE FUZZING OF COMPUTER CODE

The disclosed subject matter provides for software testing using metaphor based language fuzzing. Metaphor based language fuzzing can decompose a code segment into a metaphor representing the code segment. The metaphor can be mutated based on determined logical perturbations to any element of the metaphor. The mutation of the metaphor can act as a surrogate for mutation of the code segment. The mutated metaphor can be analyzed to reveal performance differences in comparison to the code segment. These performance difference can be correlated to mutation of the metaphor such that a corresponding mutation of the code segment can be correlated by extrapolation. Moreover, mutators can be stored and reused on other metaphors. Furthermore, employing a metaphor as a root language surrogate can facilitate generating a reduced number of mutators as compared to directly mutating code segments in a plurality of computer languages.