The importance of the relationship between the software variability and defect proneness of software modules is well recognized. The utilization of variability can strengthen the software versatile, but as the software variability increases, the software complexity can increase correspondingly. Most variability realization techniques are based on configuration, and that the variability realization code correlate with one configuration options may scatter across many software modules, which could easily induce defect and lead to dead code. This paper analyzes the preprocessor based realization of the variability, series of variability metrics are defined and the variability from different granulites is analyzed to verify whether the high variability can cause high defect. Experimental result shows that the software variability and the defect have statistically significant relationship.
[1]
Sven Apel,et al.
Type-Checking Software Product Lines - A Formal Approach
,
2008,
2008 23rd IEEE/ACM International Conference on Automated Software Engineering.
[2]
Sven Apel,et al.
An analysis of the variability in forty preprocessor-based software product lines
,
2010,
2010 ACM/IEEE 32nd International Conference on Software Engineering.
[3]
Michael D. Ernst,et al.
An Empirical Analysis of C Preprocessor Use
,
2002,
IEEE Trans. Software Eng..
[4]
Pankaj Kumar,et al.
Relationship between Crosscutting Concerns and DefectsNOTE FROM ACM: It has been determined that this paper plagiarized earlier works. Therefore ACM has shut off access to this paper.
,
2011,
SOEN.