Grammatical inference algorithms for the generation of visual languages

Abstract

A grammatical inference technique tailored for the definition of the syntax of visual languages is presented. The technique employs a functional similarity measure between icons, the primitives of visual languages. An algorithm based on this technique has been implemented and utilized for the development of the inference module of a system for the automatic generation and use of user-defined visual languages. Experimental results comparing the behaviour of the algorithm with that of Levine [18] are shown. The results suggest the efficacy of the algorithm for the generation of application-specific visual languages.