Improving industrial adoption of software engineering research
B. L. Robinson, Patrick Francis · 2010
Advances in software engineering have led to the creation of many new software engineering techniques. However, industrial adoption of these techniques is often quite low, as development organizations are skeptical of their value and applicability. Empirical studies are commonly used to show this value to potential adopters, with open source software used as an approximation of industrial applications. However, little data exists on the similarity of open source and industrial software. We present a large metrics-based study comparing the most commonly evaluated open source programs to a large set of industrial programs. Source metrics are calculated and compared between 24 open source and 21 industrial programs. The results identify open source programs that are most similar to industrial programs. Using these identified open source programs in empirical studies can provide the best generalization to industrial software.