[Block Lecture] Beyond the Numbers: Understanding and Applying Software Metrics

Prof. Dr. Horst Lichter

Research Group Software Construction
RWTH Aachen University, Germany

Software metrics have been used for decades for evaluation, and a plethora of metrics have been proposed in software engineering. However, it is increasingly apparent that metrics are often not applied in accordance with their definitions, leading to incorrect interpretations and conclusions. I have personally noticed this time and again when reviewing papers that use metrics to perform quantitative assessments.

Since software metrics are essential for evaluating software artifacts and processes, all software engineers involved in research or practical work should be familiar with their conceptual foundations. This tutorial aims to convey the essence of this knowledge.

In this tutorial, I will introduce the conceptual and theoretical foundations of software metrics, outline the essential characteristics of metrics, and explain how metrics can be classified along different dimensions. This knowledge is the fundamental prerequisite for correctly using and applying metrics. In the second part, I will discuss how metrics should be developed and what needs to be considered in this process. To illustrate the foundations and characteristics, I will introduce and discuss practical metric examples.

The tutorial will be suitable for master’s students, particularly doctoral candidates, as well as software engineering practitioners.

1. Motivation & Foundations of Software Metrics
2. Model and Measurement Theory
   • General model theory with examples
   • Representational theory of measurement
   • Scales (with examples)
   • Classification of metrics and metric properties using examples
3. Discussion about well-known software metrics
   • Code complexity metrics
   • Cohesion metrics, etc.
4. Development of Software Metrics
   • A small but not trivial metric model and a concrete metric will be developed interactively.
5. Summary and discussion