Introduction (Andreas Jedlitschka & Dietmar Pfahl)
Presentation slides

Abstract:
In Software Engineering (SE), many different areas for decision-making can be found along the software life cycle. Five major areas of decision-making research is progressing in SE, i.e. requirements, architecture and design, adaptive and corrective maintenance, project planning and control, and verification and validation. A major role of experimental software engineering research is to provide decision-makers in software development with the type of evidence that they need to support informed decision-making when introducing new methods, techniques, and tools (technologies). Systematically selecting appropriate technologies requires the identification of alternatives that, based on currently existing empirical evidence , promise to be most appropriate for the business or project goals in the context of a specific organization.

Talk on decision support in software project management

Speaker: Stefan Wagner, University of Stuttgart, Institute of Software Technology, Germany
Title: Evidence-based decision making in software engineering: Project management
Presentation slides

Abstract:
In the project management of software development and maintenance projects, the project manager needs to make a variety of decisions on very different aspects of the project. This starts with early planning decisions such as the cost budget, schedule and ressources. The manager also needs to decide which process to follow and how to tailor it for the specific project. Then he or she further details that by specifying the concrete processes and choosing methods for all parts of the development process. All these decisions need a valid basis. Evidence-based decision making can help here to ground these decisions not only on experience but on empirical evidence from old project data or from other studies. Hence, there are also two types of ways to include evidence in this decision making:
(1) the practicioners themselves collect data und analyse it to form evidence and
(2) academics run studies to analyse methods and techniques to form evidence in which contexts they are effective and efficient.
This talk will look at both types of evidence-based decision making and will discuss them using several examples, such as analysing project factors for process tailoring, analysing defects for optimising defect-detection techniques, analysing source code attributes for predicting bugs and analysing clone detection.

Talk on decision support in software quality management

Speaker: Per Runeson, Lund University, Sweden
Title: Strategic and operational decision support in quality management
Presentation slides

Abstract:
Software quality management includes several decisions, strategic and operational ones. Strategic decisions include which quality assurance activities, like inspection and testing, are conducted when, by whom, and to what extent. Other strategic trade-offs include making investment in test automation. Which testing should be automated first, and what should not be automated? In a software product line context, issues on testing the products vs. testing the platform is also a strategic issue. Operational issues include quality measurements, assessment and prediction. First and foremost, the general question of "when to stop testing" is an issue for every project manager. Prediction models exist in abundance, but which can be trusted? This mini-tutorial will address both strategic and operational decisions, based on empirical studies in software engineering. The studies comprise experiments to compare defect detection techniques, case studies to observe and learn from large-scale industrial testing practices, surveys which give an overview of the state-of-practice on quality assurance, and systematic literature reviews, which synthesize the empirical evidence in a certain field of research.