Practical Ada Mutation Testing Tool

Software testing is one of the most problematic phases of software development. This stems from the inability to gain acceptable confidence that a piece of software will not fail after deployment. Mutation testing is one form of white-box software testing that holds promise in assuring that a piece of software does not contain faults after testing is complete. Mutation testing does this by mutating the source code and then finding test cases that are able to distinguish the original code from the mutated code. This creates a set of test cases that are termed as "mutation adequate." These test cases are considered as having a greater ability to detect software faults and they are then used as the test inputs to the program. Unfortunately, mutation testing techniques have not been practical -- due to high performance costs. We propose to research these issues to prove the feasibility of developing a mutation testing tool for Ada software components. This work will yield a tool that provides NASA with software that should be less likely to contain faults after testing. This reduction in software faults will provide significant cost reductions associated with maintaining NASA software that has been tested using mutation testing.