Automated generation of consistency-achieving model editors

Abstract

The advances of domain-specific modeling languages (DSMLs) and their editors, created with modern language workbenches, have convinced domain experts of applying them as important and powerful means in their daily endeavors. Despite the fact that such editors are proficient in retaining syntactical model correctness, they present major shortages in mastering the preservation of consistency in models with elaborated language-specific constraints. Consequently, there is a demand for automating procedures to support editor users in both comprehending as well as resolving consistency violations. In this paper, we present an approach to automate the generation of advanced editing support for DSMLs offering automated validation, content-assist and quick fix capabilities beyond those created by state-of-the-art language workbenches that help domain experts in retaining and achieving the consistency of models. For validation, we show potential error causes for violated constraints, instead of only the context in which constraints are violated. The state-space explosion problem is mitigated by our approach resolving constraint violations by increasing the neighborhood scope in a three-stage process, seeking constraint repair solutions presented as quick fixes to the editor user. We illustrate and provide an initial evaluation of our approach based on an Xtext-based DSML for modeling service clusters.