Abstract
A context-sensitive graph grammar with neighborhood controlled embedding (context-sensitive NCE graph grammar) is an extension of the well-known (context-free) NCE graph grammar, and is flexible enough for use in formalizing the graph-syntax rules of a wide range of visual languages. We discuss its parsability, that is, whether a derivation for a graph generated by the grammar can be found by the reverse application of the grammar's productions. We then present a sufficient condition for determining whether a context-sensitive NCE graph grammar is parsable. We also describe a system called GRAVIS that is intended to support the design of visual languages on the basis of context-sensitive NCE graph grammars. Users can easily enter node labels, edge labels, and productions, and then derive and parse graphs with respect to the productions, by using GRAVIS's GUI.