Empirical software engineering: a roadmap report from a workshop held at STEP 2002, Montreal, October 2002

Abstract

The growing awareness of the need to employ evidence-based arguments to support the practices of software engineering, rather than arguments based upon advocacy, has led to a growing interest in the related area of empirical practices. This workshop, held at STEP 2002, set out to address two questions. What techniques do we need for empirical software engineering? What are the 'grand challenges' that empirical software engineering should be addressing as a matter of priority? For practical reasons, these were addressed in reverse order, and we began by identifying what each participant saw as being the challenges facing the wider acceptance and use of empirical techniques in software engineering. We then classified and grouped the challenges so identified under the three headings of resource, technical, and strategic. The next step involved examining how the particular challenges could be addressed, what proposals for action we would put forward, and how these proposals might be followed up. A core element we identified was the need to provide some form of centralized resource (such as a Web site) that could support both research and teaching. In this paper we outline the role that such a site could have, and also identify some of the elements that are likely to be needed in any educational module that addresses this topic.