University of Oslo
What can - and should - empirical software engineering learn from empirical studies in psychology?
Magne Jørgensen, University of Oslo
Software development is about people solving problems. They do this alone or in groups, for themselves or for others, with knowledge, experiences and biases. People solving problems have been empirically studied for more than two hundreds years in psychology, affecting not only treatment processes, but also economics, management, marketing, teaching and numerous other disciplines. What can we learn from how psychology researchers do their empirical studies and succeed in achieving useful results and affect practice? In this talk I will examine similarities and differences in use of empirical methods in psychology and software engineering and summarize this in what I argue has the potential of improving the quality and impact of our empirical software engineering studies.
Magne Jørgensen is a professor of software engineering at the University of Oslo and chief research scientist at Simula Metropolitan. He has extensive industry experience as consultant and manager and currently serves at the Norwegian digitalisation board where he advises governmental software projects. His research interests include project management, evidence-based software engineering and expert judgment. His recent book on effort estimation can be downloaded for free from: tinyurl.com/timepredictions.
NOKIA Cassandra; Use of machine learning
for further software quality improvement
Georgios Niros, NOKIA Networks
One of the biggest problems in big software companies is technical debt and cost of poor software Quality. NOKIA Cassandra, a prototype Decision Support System leveraging Machine Learning, learns to distinguish between source code of good and bad quality at procedure/method scope and gives a ranking of procedures/methods that need to be reviewed and may need to be refactored. Cassandra also provides a more granular view of what caused bad source code quality. Through its systematic use into the code commission process, Cassandra can help improve source code quality. This process is an integral part of any effort to achieve a left shift in software fault detection.
Georgios Niros is an R&D Manager for NOKIA TAS FMS Sq. Group as well the Project Lead and Concept Owner of NOKIA Cassandra. He is a technical manager with over 15 years’ experience designing and delivering telecommunication software serving different positions over several Siemens and NOKIA products. He is a Machine Learning evangelist and passionate about the benefits of AI and ML can bring to the software engineering process. Leading development teams, he is always looking ways to improve towards quality left shifting, efficiency and automation delivering quality products on time. As the concept owner of Cassandra framework, he envisioned a system able to recognize patterns by assessing passed historical data and correlating software characteristics and metrics so that to effortless classify fault prone software as well as to get proposals for refactoring actions. He is a respected technical lead and Quality advocate who regularly shares his expertise on topics of Software development, Quality and Machine Learning throughout NOKIA organization. Cassandra project is one of the few ideas got funded to operate under NOKIA internal start-up approach.