Keywords:

Keywords:
Astract: In this paper, we discuss how to make a generic model versioning system language-specific by using various adaptation techniques. In particular, we recap some lessons learned during the AMOR project and outline some open challenges for adaptable model versioning systems.

Keywords:

Keywords: ATL, UML profiles, transformation language extension
Astract: The software engineering process has significantly changed over the last decade. Whereas in the past models were only used for communication and brainstorming purposes, this philosophy has shifted drastically. Model-Driven Engineering (MDE) is the keyword that is guiding the current engineering direction. Models are the key artifact and all development steps are aligned to these models. Sophisticated modeling techniques have been invented to ensure a consistent and comprehensive technological basis. The Unified Modeling Language (UML) was introduced by the Object Management Group (OMG) to standardize and support different modeling aspects like structural, behavioral, and architectural models. The huge success of UML is not only due to the versatility of the language but also because of the highly-developed language extension mechanism in form of UML profiles. UML profiles may be defined for tailoring UML to specific domains and technological platforms.
Apart from modeling languages, the technique of model transformation plays a crucial role for the model-driven approach. Model transformations aim at transforming an existing source model into some desired target model. Different model transformation languages were invented that support various kinds of model transformations. One of them is the ATLAS Transformation Language (ATL), which is currently one of the most widely used transformation languages. While modeling languages defined by metamodels are directly supported in an ATL transformation, the use of UML profiles demands for a complex work-around. It would be desirable, however, to simplify the handling of UML profiles in such a way that profiles are, like metamodels, represented as language definitions.
The contribution of this master's thesis is to extend ATL for a native UML profile support. New language constructs are integrated into the abstract and concrete syntax of ATL to ease the use of profile-specific information within a model transformation. Apart from the extension of the ATL syntax, also an operational semantics for the new constructs is defined by translating the extended ATL to standard ATL. The goal of this work is to enrich the ATL language with new language-inherent constructs and keywords. During the implementation phase several technological possibilities as well as limitations were encountered and discussed. This collection of lessons learned can be seen as a guideline for future extensions of ATL, and as a starting point for a critical discussion about the extensibility of ATL.

Keywords:
Astract: Aspect-orientation provides a new way of modularization by clearly separating crosscutting concerns from noncrosscutting ones. While aspect-orientation originally has emerged at the programming level, it now stretches also over other development phases. There are, for example, already several proposals for Aspect-Oriented Modeling (AOM), most of them pursuing distinguished goals, providing different concepts as well as notations, and showing various levels of maturity. Consequently, there is an urgent need to provide an in-depth survey, clearly identifying commonalities and differences between current AOM approaches. Existing surveys in this area focus more on comprehensibility with respect to development phases or evaluated approaches rather than on comparability on bases of a detailed evaluation framework. This article tries to fill this gap focusing on aspect-oriented design modeling. As a prerequisite for an in-depth evaluation, a conceptual reference model is presented as the article's first contribution, centrally capturing the basic design concepts of AOM and their interrelationships in terms of a UML class diagram. Based on this conceptual reference model, an evaluation framework has been designed, resembling the second contribution, by deriving a detailed and well-defined catalogue of evaluation criteria, thereby operationalizing the conceptual reference model. This criteria catalogue is employed together with a running example in order to evaluate a carefully selected set of eight design-level AOM approaches representing the third contribution of the article. This per approach evaluation is complemented with an extensive report on lessons learned, summarizing the approaches' strengths and shortcomings.