Graphical views can be an efficient means to facilitate the understanding of complex networks (dependencies, relationships, connections, etc.). It’s relatively easy to set up a pipeline to generate a graphical view from your model or data source. But quite often such a generated view ends up in an illegible mess: connections running criss-cross, no idea where to start looking, and bad responsiveness of the application.
We will present techniques to avoid such information overflow and support users in getting the insights they need. After all, we want graphical representations to help solving our problems instead of yet another tool for generative art.
The key is to reduce information in a smart manner. We will discuss the following approaches:
- Offer domain-specific filter options and fine-tune them for the most relevant use cases
- Generate nested graphs to improve navigation and discovery through expanding and collapsing nodes
- Preserve application responsiveness with client-side level-of-detail (LOD) filtering
Finally, we will demonstrate how these concepts can be realized with Eclipse Sprotty using a client-server architecture both for web applications and for desktop tools, e.g. built on Eclipse Theia or VS Code.