Label management: Keeping complex diagrams usable

Most visual languages are not purely graphical but include textual labels to complete the picture. However, in some languages labels tend to become rather long and thereby enlarge diagrams considerably. Since today's state-of-the-art development tools usually display diagrams in full detail, users must often scroll through the diagram or zoom out until the diagram fits inside the available drawing area, but then ceases to be legible. In this paper, we address this problem by examining ways to dynamically shorten the text of labels to keep the size of a diagram manageable. We introduce a number of label shortening strategies, explain ways to integrate them into diagram generation processes based on automatic layout algorithms, and explain their relation to the established focus and context approach which aims at solving a similar problem. We evaluate our strategies based on the SCChart visual language and an open-source, Eclipse-based modeling environment.