Aus SDQ-Wiki
Zur Navigation springen Zur Suche springen
Datum 2019/06/19 11:30:00 – 2019/06/19 12:30:00
Ort Gebäude 50.34, Raum 333
Vortragende(r) Yves Schneider
Forschungsgruppe ARE
Titel An Empirical Study of Architectural Decay in Open-Source Software
Autoren Duc Le, Daniel Link, Arman Shahbazian and Nenad Medvidovic
PDF https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8417151
URL https://doi.org/10.1109/ICSA.2018.00027
BibTeX https://ieeexplore.ieee.org/xpl/downloadCitations?recordIds=8417151&download-format=download-bibtex&citations-format=citation-only
Abstract Architecture is the set of principal design decisions about a software system. In practice, new architectural decisions are added and existing ones reversed or modified throughout a system's lifetime. Frequently, these decisions deviate from the architect's well-considered intent, and software systems regularly exhibit increased architectural decay as they evolve. The manifestations of such ill-considered design decisions are seen as “architectural smells”. To date, there has been no in-depth study of the characteristics or trends involving this phenomenon. Instead, when referring to architectural smells and their negative effects, both researchers and practitioners had to rely on folklore and their personal, inherently limited experience. In this paper, we report on the systematic step we have taken in investigating the nature and impact of architectural smells. We have selected a set of representative architectural smells from literature and analyzed their instances in 421 versions from 8 open-source software systems. We have (1) developed algorithms to automatically detect instances of multiple architectural smell types, and (2) analyzed relationships between the detected smells and the lists of issues reported in the systems' respective issue trackers. Our study shows that architectural smells have tangible negative consequences in the form of implementation issues as well as code commits requiring increased maintenance effort throughout a system's lifetime.

Best Paper Award ICSA 2018