Empirical software engineering: From discipline to interdiscipline

Abstract

Empirical software engineering has received much attention in recent years and coined the shift from a more design-science-driven engineering discipline to an insight-oriented, and theory-centric one. Yet, we still face many challenges, among which some increase the need for interdisciplinary research. This is especially true for the investigation of social, cultural and human-centric aspects of software engineering. Although we can already observe an increased recognition of the need for more interdisciplinary research in (empirical) software engineering, such research configurations come with challenges barely discussed from a scientific point of view. In this position paper, we critically reflect upon the epistemological setting of empirical software engineering and elaborate its configuration as an Interdiscipline. In particular, we (1) elaborate a pragmatic view on empirical research for software engineering reflecting a cyclic process for knowledge creation, (2) motivate a path towards symmetrical interdisciplinary research, and (3) adopt five rules of thumb from other interdisciplinary collaborations in our field before concluding with new emerging challenges. This supports to elevate empirical software engineering from a developing discipline moving towards a paradigmatic stage of normal science to one that configures interdisciplinary teams and research methods symmetrically.

The Framework For The Discipline Of Software Engineering in Connection to Information Technology Discipline

“…SE is an evolving engineering discipline and covers a wide application domain like problem modeling and analysis, software design, software verification and validation, software quality, software process, software management, and many more [27]. Today SE is characterized by social, cultural and human-centric issues as it provides effective team procedures as individuals need to deal with different people at various stages of software development [5]. According to Fernandez and Passoth, those issues include application domain-specific questions (e.g., on domain-specific terminologies, concepts, and procedures), ethical questions (e.g., moral assessments in the context of safety-critical situations), juridical questions (e.g., on data privacy or regulations of algorithms and their environment respectively), psychological questions (e.g., on improvements of team communications or working environments), or social and political questions (e.g., on societal impacts of software-driven technologies, the concerns of heterogeneous actors, or accountability issues).…”

Open Science in Software Engineering

“…Hence, scientific practices often remain rooted in trust rather than being rooted in transparent scientific processes. Yet and as laid out by Mendez and Passoth [10], it is theory building which constitutes a crucial foundation to our avenue towards turning our engineering discipline into a more scientific, evidence-based one, same as it was the case for many other disciplines before. Transparency, credibility, and reproducibility are cornerstones in building and evaluating robust and reliable theories for our still emerging field and open science provides a solid foundation to achieve that goal.…”

“…These factors very often dictate in one form or the other which submissions eventually make it into the publication landscape and which do not, and which publications are cited and which are not. As a consequence, publication and citation regimes -although inherently rooted in scepticism -have also much to do with trust and convictions [10]; something which holds for most, if not all, scientific disciplines. Transparency is therefore key to break with scientific theories being grounded in common sense, taken-for-granted knowledge, hopes, convictions, and provisional beliefs.…”

Challenges in Survey Research

“…The ultimate goal of empirical software engineering is, in one way or another, to build and evaluate scientific theories by applying empirical research methods [14]. Survey research is one such means to contribute to theory development [44] as the main objective for conducting a survey is either of the following [64,52]: explorative, descriptive or exploratory.…”