Issue No. 04 - July-Aug. (2013 vol. 30)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MS.2012.111
Yang-Ming Zhu , Philips Healthcare
David Faller , Philips Healthcare
Defect density is the ratio between the number of defects and software size. Properly assessing defect density in evolutionary product development requires a strong tool and rigid process support that enables defects to be traced to the offending source code. In addition, it requires waiting for field defects after the product is deployed. To ease the calculation in practice, a proposed method approximates the lifetime number of defects against the software by the number of defects reported in a development period even if the defects are reported against previous product releases. The method uses aggregated code churn to measure the software size. It was applied to two development projects in medical imaging that involved three geographical locations (sites) with about 30 software engineers and 1.354 million lines of code in the released products. The results suggest the approach has some merits and validity, which the authors discuss in the distributed development context. The method is simple and operable and can be used by others with situations similar to ours.
Software d, Software metrics, Approximation methods, Software quaility, Analytical models, Software performance, Performance evaluation, distributed development, defect density, evolutionary development, code churn
D. Faller and Y. Zhu, "Defect-Density Assessment in Evolutionary Product Development: A Case Study in Medical Imaging," in IEEE Software, vol. 30, no. , pp. 81-87, 2013.