Our work on GenEd is especially related to designing VL editors and to theory of VLs. There exist many approaches to the theory of visual language specifications and formalisms, but the number of system implementations is very limited. Mostly, these approaches extend string grammar formalisms. A complete and recent overview is out of scope of this paper. However, we like to mention a few approaches: generalizations of attributed grammars (e.g. picture layout grammars [12]), positional grammars (e.g. [13]), graph grammars (e.g. [14,15,16]), and algebraic or type-theoretic formalisms (e.g. [17,18]). Other work closely related to our approach uses (constraint) logic or relational formalisms (e.g. [19,20,21,22,23,24,25]) for representing spatial relationships. A more detailed review of closely related work on VL theory can be found in [4].
GenEd's philosophy of a free-form general purpose editor supported by (incremental) visual parsing is in contrast to the following two approaches. Escalante [26] is an environment for the rapid construction of VL applications. It supports the construction of languages that are based on graph models (nodes and edges). The emphasis of Escalante is on user interface construction and not on VL semantics. [17] describes an approach to generate syntax-oriented visual editors for formally specified languages. It is embedded into an algebraic specification formalism but deals only with syntactic issues of VLs.
We conclude our review with two approaches that are closely related to GenEd. DiaGen [27] is a generator for diagram editors providing direct manipulation and animation of diagrams. Diagrammatic VLs can be specified by a hypergraph grammar which is used to generate a specialized editor. The extension of DiaGen by a parser supporting general editing is in progress. Another approach [28] is concerned with the automatic construction of (pen-based) user interfaces for visual languages specified by constraint multiset grammars. The VL specification is used to generate a language-specific parser. A general editor tool utilizes these parsers to control and support the editing process of corresponding VLs.
A major distinction to the approaches mentioned in this section is the focus of GenEd on the design of formal specifications of visual notations. Our experience suggests that this is an iterative process with mutual dependencies between the visual notation and its specification. GenEd supports this development cycle by offering a free-form graphic editor and object-centered specifications based on a formal theory with clear semantics. The general, built-in visual parser is directly driven by these specifications.