Software developers commonly use mailing lists to communicate and discuss issues. Some of the emails discuss architectural design decisions (ADDs) to design the components of a system, or achieve quality attributes. In this paper, we explore (a) The types of ADDs in mailing lists, and their co-occurrences, and (b) The relations between ADDs in mailing lists, and the ones in issue trackers (e.g. Jira). We applied qualitative methods, as well as deep learning classification to classify ADDs in mailing lists. Furthermore, we applied similarity algorithms to detect related ADDs between emails and issues. Our findings support empirically-grounded tools to identify and classify ADDs in mailing lists and issue trackers for future system design.

Mohamed Soliman is a Junior professor in Paderborn university since January 2024, and previously an FSE Postdoc fellow in the university of Groningen since September 2019. Before joining the university of Groningen, Mohamed made his PhD from 2013 to 2018 in the university of Hamburg with a thesis entitled “Acquiring Architecture Knowledge for Technology Design Decisions”. His main research passion and experience is software architectural knowledge capturing and reuse. In addition, Mohamed worked as software developer and technical leader for 8 years in different software companies, including multinational companies (e.g. Hewlett Packard (HP)), where he designed and implemented applications using different technologies and methods. Mohamed received his Bachelor and Master degree in Computer and Systems Engineering from Ain-Shams University, Egypt in 2005 and 2010 respectively.