The Evolution of Empirical Methods in Software Engineering

TL;DR

This paper traces the historical development of empirical methods in software engineering from the 1960s to the 2010s, highlighting methodological shifts, key contributions, and current trends in the field.

Contribution

It provides a comprehensive chronological overview of the evolution of empirical methods in software engineering, including key papers, venues, and future outlooks.

Findings

Empirical methods have evolved through five distinct iterations since the 1960s.

Current trends include increased focus on research plurality, human factors, and data synthesis.

The paper identifies future directions and ongoing challenges in empirical software engineering.

Abstract

Empirical methods like experimentation have become a powerful means to drive the field of software engineering by creating scientific evidence on software development, operation, and maintenance, but also by supporting practitioners in their decision making and learning. Today empirical methods are fully applied in software engineering. However, they have developed in several iterations since the 1960s. In this chapter we tell the history of empirical software engineering and present the evolution of empirical methods in software engineering in five iterations, i.e., (1) mid-1960s to mid-1970s, (2) mid-1970s to mid-1980s, (3) mid-1980s to end of the 1990s, (4) the 2000s, and (5) the 2010s. We present the five iterations of the development of empirical software engineering mainly from a methodological perspective and additionally take key papers, venues, and books, which are covered in chronological order in a separate section on recommended further readings, into account. We complement our presentation of the evolution of empirical software engineering by presenting the current situation and an outlook in Sect. 4 and the available books on empirical software engineering. Furthermore, based on the chapters covered in this book we discuss trends on contemporary empirical methods in software engineering related to the plurality of research methods, human factors, data collection and processing, aggregation and synthesis of evidence, and impact of software engineering research.