“Principles and a Process for Successful Industry Cooperation—the Case of TUM and Munich Re,” by Maximilian Junker and his colleagues, described the success factors behind a long-term collaboration between Technische Universität München (TUM) and the reinsurance company Munich Re. This collaboration has facilitated eight projects focused on the quality of software development artifacts. It has been beneficial both to Munich Re (in terms of improved software engineering methods) and TUM (in terms of two spin-off companies and numerous research publications).
Retrospective analysis determined that the collaboration’s success resulted from following 12 key project management principles grouped into four categories. The project content principles are to select relevant problems, select manageable-sized problems, and choose concrete, noninvasive solutions. The staffing principles are to include industry staff in project teams, involve the problem owners, and involve management. The organization principles are to maintain regular meetings and disseminate project results. The mindset principles are to allow creative leeway, enable trustful interaction, be open to criticism, and appreciate mutual interests. Although these principles might lead to general nodding and the question “So what?,” the paper compellingly explained what those principles mean, why they’re important, and the research process that helps realize them. This paper was part of the ICSE ’15 Software Engineering Research and In dustrial Practice Workshop and received the IEEE Software Best Paper Award (see p. 4.); access it at http:// goo.gl/hK1yCk.
“Fast Feedback Cycles in Empirical Software Engineering Research,” by Antonio Vetrò and his colleagues, described the ongoing discussion on Empirical Software Engineering 2.0 as a way to improve empirical research results’ impact on industrial practice. A fast feedback cycle is enabled by first identifying key concepts that need studying and then designing and quickly executing a small proof-of-concept study to demonstrate the approach’s potential benefits, by leveraging the use of interactive data analysis and mining.
In an illustrative example, Vetrò and his colleagues worked with Scrum-based projects to develop and evaluate a mechanism that analyzes user stories to discover scope creeps regarding project goals. The authors used their model to infer topics and then received quick feedback from the industrial partner. This fast feed
back let the researchers successfully tune the extraction model.
One study participant said, “I found it interesting to see, in the pilot study, how clustering the words of our user stories pointed out what underlying interdependencies we have in our requirements. We are now applying the same principles of the iterative feedback methodology to improve our processes, for example, concerning estimation quality and semantic clarity of requirement specifi cations.”