Workspace updates of visual models

Abstract

In MDE, large models must be collaboratively developed in teams. Collaboration is usually supported by optimistic versioning based on workspaces and a repository. Workspace copies of a model are synchronized with the repository version by an update function. Most update functions currently available compromise the consistency of a model. We present a new approach which guarantees updated models to be consistent and processable by standard visual editors without post-processing. Our approach assumes changes in the repository to be represented as a consistency-preserving edit script. This edit script is executed in a controlled, interactive manner on the workspace copy. Unlike 3-way merging, our approach needs no conflict analysis. Conflicts are detected and resolved interactively. Our approach offers improved usability and requires less implementation effort than existing approaches.

Automatic Inference of Rule-Based Specifications of Complex In-place Model Transformations

“…[24], and other kinds of complex change operations such as recurring and highly schematic editing patterns [28]. That means, language-or project-specific change operations can be specified by model transformation rules which can be used to adapt a variety of MDE tools supporting model evolution; advanced model editors [28], modern refactoring tools, high-level differencing [20] and merging tools [21], or evolution analysis tools [16] being examples of this. However, generic model transformation techniques and tools supporting inplace transformations are commonly based on the abstract syntax of modelling languages, Henshin [5] and VIATRA2 [7] being examples of this.…”