Main Content

Metamodeling and Graph Grammars

Funded by the German Research Foundation (2012 - 2020).

Visual modeling languages play an essential role in understanding and constructing systems, both hardware and software systems in computer science and systems of other kinds such as production systems. In model-driven software development, models even become the central artifacts of software development. Depending on the purpose of modeling, different modeling languages, general and domain-specific as well as visual and textual, are needed.
Current project

Generation of development environments for modeling languages
The goal of the project is to increase the level of automation for domain-specific modeling languages (DSMLs). Since tool development seems to be the biggest obstacle for the use of DSMLs, we want to identify the core problems and develop new concepts, methods and meta-tools in order to accelerate and simplify tool development for DSMLs. To this end, we investigate the specification of DSML-specific tool environments at a higher level of abstraction to take advantage of model-based development. Here, the metamodel of a DSML represents the primary source material together with a highly abstracted tool specification. This information is sufficient for the language-specific meta-tool generation of the IDE tools. The DIE tools generated in this way enable the generation, completion and repair of models of the respective metamodel and thus always result in valid models.

Project Collaborators

Nebras Nassar, M. Sc. (Philipps University Marburg)

Previous project

Integration of two paradigms for the definition of visual modeling languages
In this project, we consider two complementary approaches to defining visual modeling languages: Metamodeling and Graph Grammars. While metamodeling represents a declarative language design, graph grammars define a language in a constructive way. To better exploit the advantages of these two paradigms, we want to integrate them appropriately. Metamodels shall be translated into equivalent model grammars that allow for a well-founded automatic generation of instance models. The formal basis is the theory of algebraic graph transformation. We intend to use this metamodel translation for the development of user-friendly model editors on the one hand and for the systematic testing of model transformations on the other hand. The new methods will be implemented on the basis of the Eclipse M odeling project and evaluated on two reference applications.

Project collaborators

Dr. Thorsten Arendt (Philipps University Marburg)
Dr. Hendrik Radke (University of Oldenburg)

Cooperation partners

Prof. Dr. Annegret Habel (University of Oldenburg)

Auxiliary activities

Within the project we are looking for students to implement a generator for instance models based on the developed concepts. The generator is to be created as an Eclipse plugin based on the Eclipse modeling project. If you are interested in assisting (basis: 30 or 40 hours/month) with the mentioned topic, please contact us.