Volume 13, Number 3

Inherent Quality Metrics for Continuous Software Quality Enhancement

  Authors

Ning Luo and Linlin Zhang, Intel Asia-Pacific Research & Development Ltd, China

  Abstract

Traditional software quality metrics based on bug number and pass rate can only provide us afterthought post validation & product release. In this paper, we propose some new inherent software quality metrics for proactive quality control during development phase, including Lines of Code (LOC#), Cyclomatic Complexity and Code Churn. In this paper, citing one ultra-large-scale software - Intel Media Driver as one example, we introduce the reason to choose those metrics, our experience on leveraging those metrics to improve the software quality, and our turn-key solution for automatic metrics data collection and analysis. We expect the identified metrics can help more researchers to form the corresponding research agendas and the experiences sharing can help following practitioners to apply similar enhancements.

  Keywords

Perceived Software Quality Metrics, Inherent Software Quality Metrics.