Fine-Grained Semantic Metrics for Object-Oriented Software

摘要

Software metrics help developers and managers assess the quality of their software. Whereas traditional metrics count aspects of code's syntax, semantic metrics quantify the meaning of the code's functionality in the domain. Because semantic metrics do not depend on the structure of code, they can be calculated from documentation before the code has been written. Existing semantic metrics are computed at the system level and class level and depend on having a conceptual graph-based knowledge base. This paper presents a new suite of semantic metrics that are computed at a finer grain and that can be used with a wider range of knowledge bases.