Text

  • Study location Room Lambda and via Zoom
Date
  • 2020-11-20 13:30–15:00

The public defense of Robbert Jongeling’s licentiate thesis in Computer Science and Engineering

The public defense of Robbert Jongeling’s licentiate thesis in Computer Science and Engineering will take place at Mälardalen University, room Lambda (Västerås Campus) and virtually on Zoom, at 13.30 on November 20, 2020.

The title of the thesis is “Advancing continuous model-based development in industry”.

The faculty examiner is Professor Manuel Wimmer, Johannes Kepler University, Linz (Austria), and the examining committee consists of:

  • Professor Tomas Bures, Charles University, Prague (Czech Republic)
  • Associate Professor Regina Hebig, Chalmers Tekniska Högskola, Gothenburg (Sweden), and
  • Professor Manuel Wimmer, Johannes Kepler University, Linz (Austria)

Reserve: Professor Sasikumar Punnekkat, Mälardalen University

The doctoral thesis has serial number 297.

Abstract

It is said that nowadays most companies are software companies. For example, innovations making services more data-driven, making devices more connected or energy efficient, and enhancing machines with autonomous functionalities, are primarily achieved through the development of new software. It is therefore very valuable to study methods of engineering
software in ways that are time effective and hence also cost effective.

One of those ways is the use of models to express abstract designs of various aspects of a system before implementing it. A unique property of software engineering is that these abstract descriptions can be made repeatedly more specific until the software itself emerges.

Moreover, some of these refinements can be performed automatically. Within this modelbased paradigm, multiple models describe the same system and can each be subject to new changes. This is pushed to the extreme in continuous development methods, where every day a working, albeit incomplete, version of the software is created. In our research, we have collaborated with industrial partners to identify challenges to making
their model-based software development of complex embedded systems more continuous. An emerging challenge within these development settings is to keep the different representations consistent that is, to make sure that the models do not contradict each other. We aim at providing approaches that address these problems in industrial practice. This has resulted in two concrete approaches that are presented in this thesis.

The first approach provides consistency checks that warn developers upon the introduction of inconsistencies between their models. The checks are lightweight since they require minimal effort to define and maintain. The second approach provides support for consistency across different models in software product lines. In such settings, models, or parts of models, are reused to describe different product variants. Changes to such models may then need to be
propagated to all the locations where that model is reused. We propose an approach that can partially automate this process and alleviate the required manual effort by domain experts. Furthermore, the approach may be used to organize the reuse of parts of models in a more structured way.

In conclusion, our contributions are aimed at advancing current model-based development practices in industry, such that they can adopt shorter development cycles and thereby increase the productivity of software development.