Toward Intelligent Drawing Constraints

Diagrammatic problem solving in domains such as architecture and design often involves continual redrawing of existing figures. Drawing programs, such as Visio and CorelDraw, accommodate redrawing using secondary correction techniques. These techniques act when a visual element is moved or changed, making additional changes to the drawing that maintain key preexisting visual relationships. These techniques include low-level constraints, such as snap and glue, and domain-specific glyphs with built-in constraints. As useful as these secondary correction techniques are, they sometimes operate in unintuitive ways, interfering with the flow of problem-solving. Here, we describe our plans to increase the effectiveness of secondary correction by making such constraints congruent with the place vocabulary of the domain in question. We introduce the idea of place vocabulary constraints (PVCs) that translate a particular place vocabulary into a set of geometric constraints. We argue that PVCs may provide a better mechanism for secondary correction during redrawing because multiple geometric constraints can be handled by a single place vocabulary constraint, because place vocabulary constraints are more congruent with users’ expectations, and because they provide a better mechanism for explaining such constraints to the user.