An Investigation of Object-Oriented and Code-Size Metrics as Dead Code Predictors

In case of biological organisms, death is unambiguous. It is not the same in software engineering, where death assumes a different meaning. For example, this term can be used to indicate source code that can be removed being it unnecessary for a given software system. In this paper, the early achievements of an empirical investigation on software metrics as predictors for dead code methods in object-oriented (OO) source code have been presented. OO design metrics and traditional code-size metrics have been considered. Data on one open source software system implemented in Java have been collected. The results suggest that five of these metrics can be used as indicators for dead code methods in OO software systems.