The Measurement of Software Product Line Variability and Its Relationship with Software Defects
Kunming Nie, Li Zhang · International Conference on Quality Software · 2012
Development of software product lines relies heavily on the use of variability, and implementation of variability is not a trivial task. Preprocessor is an effective and quite simple way to realize variability, however, it lacks structure, makes programs hard to maintain and makes programs error-prone. Although preprocessor based variability can influence program comprehension, the relationship between software defects and the preprocessor based variability is not particularly clear. This paper statistically analyses preprocessor based software variability and software defects to verify whether high variability will increase software defect count. This paper defines a series of variability metrics and analyses the variability from different granularities to verify whether the high variability can cause high defect counts. Experimental results show that the software variability and its defects have statistically significant relationships.