A constraint-based framework for diagrammatic reasoning
Bernd E. Meyer · Applied Artificial Intelligence · 2000
In the near future, visual and diagrammatic human-computer interfaces, such as those of UML-based CASE tools, will be required to offer a much more intelligent behavior than just editing. Yet there is very little formal support and there are almost no tools available for the construction of intelligent diagrammatic environments. The present paper introduces a constraint-based formalism for the specification and implementation of complex diagrammatic environments. We start from grammar-based definitions of diagrammatic languages and show how a constraint solver for diagram recognition and interpretation can automatically be constructed from such grammars. In a second step, the capabilities of these solvers are extended by allowing to axiomatise formal diagrammatic systems so that they can be regarded as a new constraint domain. The ultimate aim of this schema is to establish a constraint logic language for diagrammatic reasoning applications.