A metamodeling approach to incremental model changes

Meij, D.P.L. van der (2009) A metamodeling approach to incremental model changes.

Abstract:For the next step in raising the level of abstraction in software development, Kent (Kent, 2002) proposed an approach for Model Driven Engineering (MDE). With MDE it is possible to raise the level of abstraction such that decisions can be made regardless the implementation language. Business rules can be defined on a higher level so in fact programming happens at the level of models. Model transformations allow models to be turned into other models, for instance models of source code or other less abstract models. This thesis is about model‐to‐model and model‐to‐text transformations. Usually modelto‐ text aspects of thoughts are omitted because the text to be generated can also be modeled and thus a model‐to‐model transformation should be sufficient. Because in this thesis we focus on incremental model changes and also incremental changing the files in the file system, we explicitly studied model‐to‐text transformation approaches to keep model and text synchronized. In this thesis we use different terms for model‐tomodel and model‐to‐text transformations. We use the term transformation when we are dealing with models as logical entities, we use the term generation when we are dealing with files as physical entities. For instance, in a model‐to‐text transformation the model will be transformed and the output will be generated files; the generation of files is due to the transformation of a model. We studied the possibilities to perform specific file generation and file modification based on a set of changes to a model. In our approach we started investigating state of the art modeling environments and concluded that they were not able to perform an incremental model place, with in‐place updates to a model. Also they did not support model‐to‐text transformations so for that reason we presented our own approach. The approach we propose in this thesis is an imperative approach that is able to perform model‐to‐model and model‐to‐text transformations in parallel. We also present several implementations of our approach and benchmarked them against each other to see which implementation is the most efficient when performing different types of (straightforward) transformations. Our results show that incremental model changes are indeed possible and also in an efficient way, optimized to the target platform.
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science MSc (60300)
Link to this item:http://purl.utwente.nl/essays/59158
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page