Keywords:
Astract: Model transformation languages, the cornerstone of Model- Driven Engineering, often lack mechanisms for abstraction, reuse and debugging. We propose a model transformation framework providing different abstraction levels together with an extensible library of predefined transformations and a dedicated runtime model in terms of Coloured Petri Nets for transformation execution and debugging.

Keywords:
Astract: Model-Driven Engineering (MDE) places models as firstclass artifacts throughout the software lifecycle requiring the availability of proper transformation languages. Most of today's approaches use declarative rules to specify a mapping between source and target models which is then executed by a transformation engine. Transformation engines, however, most often hide the operational semantics of the mapping and operate on a considerable lower level of abstraction, thus hampering debugging. To tackle these limitations we propose a framework called TROPIC (Transformations on Petri Nets in Color) providing a DSL on top of Colored Petri Nets (CPNs) to specify, simulate, and formally verify model transformations. The formal underpinnings of CPNs enables simulation and veri fication of model transformations. By exploring the constructed state space of CPNs we show how prede fined behavioral properties as well as custom state space functions can be applied for observing and tracking origins of errors during debugging.

Keywords:
Astract: In the Model-Driven Architecture (MDA) paradigm the Query/View/Transformation (QVT) standard plays a vital role for model transformations. Especially the high-level declarative QVT Relations language, however, has not yet gained widespread use in practice. This is not least due to missing tool support in general and inadequate debugging support in particular. Transformation engines interpreting QVT Relations operate on a low level of abstraction, hide the operational semantics of a transformation and scatter metamodels, models, QVT code, and trace information across different artifacts. We therefore propose a model-based debugger representing QVT Relations on bases of TROPIC, a model transformation language utilizing a variant of Colored Petri Nets (CPNs). As a prerequisite for convenient debugging, TROPIC provides a homogeneous view on all artifacts of a transformation on basis of a single formalism. Besides that, this formalism also provides a runtime model, thus making the afore hidden operational semantics of the transformation explicit. Using an explicit runtime model allows to employ model-based techniques for debugging, e.g., using the Object Constraint Language (OCL) for simply defining breakpoints and querying the execution state of a transformation.

Keywords: Business Process Modelling, Goals, Measures, Languages, Variabilities, UML, Profile, Metamodel
Astract: Over the last decade more and more companies started to optimize their business processes in a way to meet its business goals. They develop business process mod- els defining which activities have to be executed in which order under which con- ditions by whom and by using which resources. For this purpose a lot of different approaches to business process modelling have been developed, which resulted in many different Business Process Modelling Languages (BPMLs). The definition of a business process has to cover many different aspects (e.g. con- trol flow, organizational view, data view, etc.). A perfect business process modelling approach would address all the different aspects. Unfortunately, none of the exist- ing approaches provides concepts for addressing all of these aspects.
Each of them concentrates on some aspects. The focus on certain aspects is mainly due to the different applications areas, e.g. business engineering or software engineering etc. Although BPMLs are well established in industry and science, a comprehensive evaluation or a framework for an evaluation to compare the different BPMLs is still missing. Thus, it is the goal of this thesis to provide an evaluation framework for the comparison of BPMLs and to apply this framework in the evaluation of the currently most popular BPMLs. The resulting framework is based on a generic metamodel that captures all of the concepts appearing in any of the state-of-the-art BPMLs. On a high level this framework addresses the following views:
Business Process Context Perspective, Behavioural Perspective, Functional Perspective, In- formational Perspective, and Organisational Perspective. An evaluation based on this framework checks whether the certain aspects in each of these perspectives is supported by the concepts of each of the considered BPMLs. In the evaluation of this thesis, we used the following languages: UML 2 Activity Diagram, Business Process Modelling Notation, Event Driven Process Chain, IDEF3, Petri Net, Role Activity Diagram. According to the evaluation we were able to identify three main problems in current BPMLs. The first problem is that the definition of the dependency be- tween business processes and their supporting software systems is inadequately supported. In our approach we support the elicitation of requirements from busi- ness process models for the software systems to be developed by extending current BPMLs with software requirements and components to ensure a business-goal ori- ented software development. The second problem concerns the variability of similar, but well-distinguished software products within a software product line. These software products not only differ in its structural definition, but also in the process to create them. Today, vari- ability modelling is a domain specific modelling technique that is limited to the structural definition of similar software products. In our approach we extend the concepts of variability modeling to integrate the dynamical aspects into the UML. The resulting approach is based on a well defined dependency between UML class diagrams and UML activity diagrams. The third problem is that current conceptual BPMLs do not provide explicit mod- elling means for process goals and their measures. The modelling of goals and its monitoring is a critical step in business process modeling. Hence, we extend the metamodels of UML 2 AD, EPC and BPMN with business process goals and per- formance measures. These concepts become explicitly visible in the corresponding models. Furthermore, a mapping of the performance measures onto the Business Process Execution Language (BPEL) enables their monitoring in an execution envi- ronment.

Keywords: aspectWebML, weaving, crosscutting concerns, modeling, model-driven softwaredevelopment, aspect-orientation
Astract: Ubiquitous web applications follow the anytime/anywhere/anymedia paradigm, meaning they are individually adapted to the context in which they are used, such as time, location and device. The implementation of such customizations are often complex, since a single change often does not only affect a specific part of the web application but instead a number of locations spread across the various levels of a web application, namely content, hypertext, and presentation. In the domain of model-driven software engineering, most web modeling languages do not take this peculiarity of customizations into account. Typically customization functionality is modeled on top of the core model and directly introduced to a number of locations. This approach however increases complexity of maintenance and extensibility and thus leads to an inefficient development process. In aspectWebML, an extension to the web modeling language WebML, customizations are instead seen as so-called crosscutting concerns and treated with the help of special aspect-oriented concepts. These allow for a clear separation of core functionality and crosscutting concerns in models, whereby a reduction in complexity of the models, an improvement in maintainability, as well as reusability of individual customizations can be accomplished.
Nevertheless, as part of the modeling process this clear separation of core functionality and crosscutting concerns has to be met with an adequate mechanism which integrates the separated parts again into a whole model to reach the result of the modeling process.
The goal of this work is therefore the implementation of an algorithm to integrate crosscutting concerns in aspectWebML on the basis of the existing aspect-oriented concepts. The requirements for such an integration will be analyzed and an appropriate technology will be chosen to implement the 13 different possibilities for modeling of customization scenarios in aspectWebML. The algorithm is then integrated into an existing modeling toolkit, which was created using EMF (Eclipse Modeling Framework) and enhanced with various functionalities to support modeling and will then be tested against a number of different customization scenarios in a case study.