RacePoint: A Software Defect Complexity Measurement Framework for Managing Data Race Bugs of Multithreaded Software
DescriptionThe project intends to improve the standard of software testing in data race detection research by establishing a trustworthy and accepted data race detection and measurement framework. Inspired by the method Function Point Analysis (FPA) as a standard size measure for software project effort estimation that provides a direct estimate of the functional user requirements of the software, which can be classified into one of five data transaction types: inputs, outputs, inquires, internal files, and external interfaces. Once the function is identified and classified into a specific type, it can be further assessed for complexity and assigned a function point measure for size. Given the availability of an objective measure of size, software developers would be able to derive effort estimates and prioritize tasks using a number of methodologies commonly practiced in software effort estimation research. Instead, the proposed work in the study will develop a measurement framework to classify the complexity of the data race bug based on the outputs of data detectors utilizing different data race detection algorithms available, and it uses detected patterns of read-write, write-write and write-read data races, together with other characteristics of data race bugs to provide a complete understanding on the corresponding bug fixing effort requirement, which provides an objective measure for the data race bugs identified.
|Effective start/end date||1/09/16 → …|