Master's Thesis Leo Knoll

Abstract

Many modeling toolchains are currently in the transition process from desktop-based to cloud-based solutions. While most versioning use cases such as branching can be efficiently covered by systems such as Git, even in the cloud, graphical models pose specific requirements to comparison and merging. This is due to the graph-based nature of models. To ensure comprehensiveness and model integrity, a text-based comparison and merge is not suitable. Moreover, when dealing also with graphical diagrams that represent underlying models, additional challenges arise, as the differences identified during a comparison have to be visualized graphically to the user so they are comprehensible.

In this thesis we provide an overview of existing model versioning techniques available on the desktop. We research how existing approaches can be applied in a cloud context and which adaptations are necessary. Unleashing the potential of modern UI technologies available in cloud-based modeling tools, we further explore the best way to visualizing changes applied to diagrams. Finally, we provide proof of concept in the form of a graphical modeling editor which implements our solution and conduct a case study to evaluate the benefits and limitations of our model versioning framework.