On the Pragmatics of Graphical Modeling
Graphical modeling employs diagrams to present models and their properties. These graphical depictions add the second dimension to one-dimensional text. Thus intrinsic properties of the diagrams, such as graphical spatial inclusion, can be used to visualize corresponding model relationships in a natural way, such as containment relations. Therefore graphical models are usually more direct while text—including formulas—tends to have more indirect features, meaning the used symbols require prior definition and explanation. If used correctly, graphical models can be easier understood by humans than textual ones. They can be used as a common ground for information exchange and discussion for different domain experts.However, the second dimension introduces new complexity. It causes users to waste a lot of time with drawing tools using archaic interaction mechanisms where they have to manually place graphical items on the canvas. Creating and maintaining of diagrams is very effort prone. Additionally it is difficult to navigate in complex models. They are either very big or have a complex nesting structure and are thus difficult to handle with the usual navigation and visualization techniques. In this thesis I present an integrated concept on handling graphical models. It comprises technological stepping-stones that each raise the level of abstraction of the user interaction with graphical models in order to concentrate on the semantic models. A key enabler for improved designer productivity is the automatic layout of diagrams. It gets employed consistently in order to free the user from the burden of manual placing and routing. Additionally it allows to synthesize multiple graphical views onto the same model easily and quickly. Thus views can be dynamically changed to the user’s needs. The user can be focused automatically to the model elements that are currently most important. The core of this approach is a view management that uses configurable conditions to synthesize new model views interactively. I present several use cases in model-driven engineering that benefit from this approach. This includes structure-based editing and dynamic focus & context views during simulation runs. For validation and to foster dissemination, several example editors with such capabilities have been developed in the Kiel Integrated Environment for Layout Eclipse Rich Client (KIELER), an Eclipse-based modeling environment available as open source.
Use and reproduction:No Creative Commons license - The provisions of the German Copyright Act (UrhG) apply.Please note that individual components of the publication may be subject to other licensing or copyright conditions.