Software Complexity as a Predictor of Defects
Date
1996-07
Authors
Benson, Jennie
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Software inspections are generally considered to be an effective means of finding and correcting errors at the coding stage of the software development cycle. Due to a number of reasons, project teams often find it impractical to inspect every module. When only a portion of the modules can be inspected, an objective criteria for selecting which modules are more likely to have errors, and therefore should be inspected, is desired.
High software complexity is contributing factor to an increased rate of errors in the software modules. Selection of a complexity metric that can be easily measured and which can be shown to reliably predict the probability of errors could be used as a criterion for selecting modules for inspection.
A tool for measuring software complexity is developed. It measures the Software Science metric of difficulty which was first proposed by Maurice Halstead. This tool was used to obtain data from a database of commercial software. The results, and what they show about the relationship of Software Science's metric for difficulty and the probability of selecting modules with errors, are presented.
Description
Keywords
computer software, defects, sampling, computer programming, quality control, computer industry
Citation
Benson, J. (1996). Software complexity as a predictor of defects (Unpublished thesis). Southwest Texas State University, San Marcos, Texas.