Abstract
One major challenge with graph-based visual languages is managing the complexity and maintaining a good readability as the density of edges in the graph increases. To improve the graph readability we propose a solution that manages the visual complexity of graph-based diagrams with many hundreds, and even thousands of nodes. It consists in applying heuristics to choose the best path for an edge and attributing to it a color defined automatically according to its origin and destination nodes. A grid system where certain areas are reserved for the display of nodes and others for the edges is used for the layout of the nodes and edges.