Keywords:
Astract: The software engineering process has significantly changed over the last decade. The paradigm Model-Driven Development (MDD) gained more and more popularity. In the past, models were typically used for brainstorming and communication purposes. This viewpoint has shifted dramatically. The software engineering process becomes more model-centric and less code-centric. Thus, models are the key artefacts in the development process and all steps rely on these models and their correctness. Sophisticated modelling techniques have been invented to ensure a consistent and comprehensive technology base. The standard for modelling structural and behavioural aspects of systems, the Unified Modelling Language (UML) was introduced by the Object Management Group (OMG). The problem is that an UML model is not necessary expressive enough to provide all complex aspects of a system. Some aspects of the domain are more expressed in natural language. Practise has shown that such situations will lead to ambiguities and to errors. Therefore the Object Constraint Language (OCL) can be used to specify conditions on models that must hold for the system. Apart from the development of standards for modelling systems, several vendors developed Computer-Aided Software Engineering (CASE) tools to provide a wide range of instruments to support the MDD approach. The support of the sophisticated techniques depends on the vendor-s realization of the standards, which is different from tool to tool. In general there is a lack of validating models in CASE tools, which is essential in the early design phase. Additionally, there is only little support for writing OCL expressions for models. Furthermore, models defined by the users of the tool are not validated against the well-formedness rules (WFR) - described with OCL - of the UML specification. An automatic validation of the OCL expressions of the UML specification before code generation allows the detection of errors in the early design phase and reduces problems for the further progress of projects. Thus, it is a huge value that CASE tools support the definition and validation of OCL expressions. The contribution of this master-s thesis is to provide an adaptive solution for managing OCL transformation in different environments or CASE tools, called ADOCLE. Therefore mapping patterns are defined to manage the schema matching between selectable environments (source and target schema). The OCL expressions are applied to a selectable source schema, which is mapped to a target schema. The goal of this work is to analyse OCL expressions and generate semantically equivalent expressions in the target schema depending on predefined transformation patterns.

Keywords:
Astract: For several years, Model Driven Software Development (MDSD) is on its rise and software models are not only used for design purposes but became a major element in the software engineering process. Modeling languages are available to describe models which are the basis for further development. In addition to the modeling languages, model transformations are an important ingredient of model driven software development. Basically, the goal of a model transformation is to transform a given source model to a desired target model. Often the model transformation fails and the troubleshooting is lengthy and timeconsuming because the mistakes can be in the transformation itself or in the metamodels. Another problem is to obtain a good test coverage. The aim of this work is to find an automated process which can test rule-based transformations. A tool is to be developed in Java, which is based on the meta language Ecore and model transformations in Atlas Transformation Language (ATL). One of the goals of the thesis is to find an automated white-box testing approach for model transformations in ATL. By the automatic generation of test instances a high test coverage should be ensured. In addition, various errors and especially runtime errors in the transformation should be detected. Beside the necessary basics such as Answer Set Programming (ASP) and testing in general the concrete implementation of the approach, the White-Box Testing (WBT) Tool, is described. Based on a collection of model transformations of the course "Model Engineering" of the Technical University of Vienna from the years 2009 until 2012 an evaluation is carried out and discussed. The method of evaluation is a case study with an objective test data set as the transformations are not created by ourselves. This set is significant with about 160 transformations. The results of the case study show that the developed WBT-Tool finds different errors. For example, failures during the validation of the models are detected and the tool finds runtime errors in the model transformation. An essential advantage is that there is no interaction except the initial setting and therefore more complex test data can be tested automatically.

Keywords: Enterprise Mashups, Mashups, Mashup Frameworks, Mashup Teaching Case
Astract: The usage of Enterprise Mashups enables users to create ad hoc applications without the need of technical knowledge. These applications can use and combine data sources of different systems. The integration of heterogeneous data in a new application is still a rather unused potential within a company.
Many companies work with a highly integrated ERP system, which contains the bigger part of the data to control the business processes. However not all necessary data is covered and accordingly not all processes can be controlled by the ERP system. Therefore the situation arises that information from various systems must be consolidated to make a decision. The usage of an ad hoc Mashup application, which is created according to necessary claims, leads to an important added value for the departments.
In line with this diploma thesis special requirements like data quality, the raised security of corporate sensitive information and the integration of legacy applications into Enterprise Mashups will be proved. Moreover a prototype for an Enterprise Mashup is implemented.
The prototype shows the combination of data from the ERP system Microsoft Dynamics AX in conjunction with other data sources. To put it into practice a solution to use the ERP system as a Mashup data source is developed. Available Mashup frameworks have been evaluated and the one with the best correlation according to the requirements has been chosen for the prototype.
The implemented Mashup shows the fast and straightforward combination of data from various applications. The prototype illustrates a teaching case for companies which want to supply their employees with the possibility to use existing data sources in the context of an Enterprise Mashup. User obtain the ability to create themselves instruments on demand for the daily business without programming skills.

Keywords: ERP Design, ERP Comparison, Software development
Astract: In today's business world, enterprise resource planning (ERP) systems are essential for larger enterprises to conduct their business.
ERP solutions fulfill several tasks in companies, including the management of resources, sales, warehouse, employees, and accounting.
Since even small and medium enterprises (SMEs) can gain profits using them, ERP solutions especially for those are developed. The main difference between them is the reduced complexity of ERP software for SMEs. Since about 98% of all Austrian companies have less than 50 employees, the market for appropriate ERP systems is large. Those statistics also apply to the entire European Union. Many SMEs do not work with an ERP system yet, but instead use Microsoft Excel and send business documents via email.
Another reason for a basic ERP system is that e-Invoicing is becoming more and more essential for SMEs. Larger enterprises are already using e-Invoicing for many years, but smaller enterprises lack the technical requirements. Most electronic data interchange (EDI) standards are too complex and the software is too expensive. Therefore a cheap and easy to use approach for SMEs to exchange business documents is desirable. The Vienna University of Technology project "`E-business Registry Permitting Enterprise Liaisons"' (ERPEL) fulfills those requirements. In order to motivate even small companies without any ERP system in use to partake, a basic ERP system is needed. A self-made solution is built especially including messaging, instead of adopting an existing open source ERP system.

Keywords: B2B e-commerce, Business Modeling, Business Process Modeling, REA, e3-value, UMM, B2B, DSL
Astract: Die heutige E-Commerce-Landschaft ist geprägt durch ständigen Wandel und Innovation. Vor allem im Umfeld des sich rasch verändernden Business-to-Business (B2B) E-Commerce wird Flexibilität im Umgang mit Änderungen immer wichtiger. Kontinuierlich wandelnde Verhältnisse zwingen Unternehmen immer häufiger dazu ihre geschäftlichen Ziele und Strategien den jeweiligen Umständen anzupassen. Dies betrifft nicht nur betriebswirtschaftlich relevante Unternehmensbereiche, sondern hat auch unmittelbare Auswirkung auf die darunterliegende technische Infrastruktur, welche die zugrundeliegenden Geschäftsprozesse zur Realisierung der Zielvorgaben umsetzt. Um den geschäftlichen Anforderungen gerecht zu werden, müssen daher auch die IT-gestützten Prozesse angepasst und berücksichtigt werden.
Die vorliegende Arbeit beschäftigt sich vor allem mit der Fragestellung, wie geschäftliche und technische Anforderungen in Bezug auf interorganisationale Informationssysteme vernetzt und integriert werden können. Um dieses Ziel zu erreichen wurde im Zuge dieser Dissertation eine Methode entwickelt, die Geschäftsmodelle und Geschäftsprozessmodelle in einer Form verbindet, die eine semiautomatische Generierung von ausführbaren Artefakten (z.B., Web Services oder Workflowspezifikationen), die konsequenterweise auf den geschäftlichen Anforderungen basieren, ermöglicht. Um die jeweiligen Problemstellungen der unterschiedlichen Sichtweisen adäquat behandeln zu können, sieht die Methode vor etablierte Techniken aus den Forschungsdisziplinen Geschäftsmodellierung und Geschäftsprozessmanagement anzuwenden. Konkret sind dies die e3-value Geschäftsmodellontologie, die Resource-Event-Agent (REA) Ontologie und die UN/CEFACT's Modeling Methodology (UMM). Des Weiteren wurde, um die genannten Ansätze integrieren zu können, ein auf semantischer Ähnlichkeit basiertes Mappingverfahren entwickelt, mit dessen Hilfe es ermöglicht wird verwandte Konzepte der unterschiedlichen Sprachen miteinander zu verknüpfen. Die entwickelte Methode wurde anhand einer konkreten Toolimplementierung evaluiert. Dabei wurden sowohl die Geschäfts- als auch die Geschäftsprozessebene als domänenspezifische Sprache implementiert. Die ursprünglichen Techniken wurden dabei in Hinblick auf deren Notation weiterentwickelt. Durch die sich so ergebende einheitliche formale Darstellung war es überdies möglich einen semiautomatischen Transformationsmechanismus zu realisieren, der auf dem zuvor erwähnten Mappingverfahren basiert. Des Weiteren wurde die Methode als auch die entwickelte Systemimplementierung gemeinsam mit Industriepartnern aus dem Printmedienbereich auf ihre Anwendbarkeit und Verständlichkeit hin untersucht.

Keywords: EDI, penetration, diversity, empirical analysis, Electronic Data Interchange
Astract: Electronic Data Interchange (EDI) has been used for decades for automated business-to-business (B2B) trading of goods and services between companies. However, EDI is still one of the key challenges of today's businesses. There are several reasons for EDI adoption or EDI non-adoption. While EDI leads to many benefits for a firm, also various drawbacks come into existence. A large variety of delimiter-based and eXtensible Markup Language (XML)-based EDI standards makes it difficult for enterprises to decide which standard to use. This decision is influenced by several factors, such as the company's industry sector or the company's trading partners. In addition, the used Enterprise Resource Planning (ERP) system plays a relevant part of well integrated EDI.
The purpose of this thesis was to examine real-world phenomena of EDI that appear in organizations in Austria. The research question this thesis aims to solve focuses on the diversity and penetration of EDI, including reasons for and against EDI adoption, EDI's share on business transactions, and the adaption between the company's internal operational structure, the ERP system, and EDI.
The findings of this thesis are based on statistical analyses of surveyed data from an empirical study of 227 companies in Austria. Of these, 103 companies used EDI, which represents an EDI penetration of 45.4% across all participants. 75.2% of the companies that used an ERP system also adopted EDI.
The results indicate that the size and the ERP system of a company have an influence on the EDI penetration. Moreover, it was found that the use of either a delimiter-based EDI standard or an XML-based EDI standards was influenced by the geographical size of the company's supplied market, the ERP system, the number of EDI partners, the number of sent EDI messages, and competitive reasons.
Furthermore, the results of the study revealed that EDI non-adopters had rather strategic than technical or financial concerns about EDI introduction. Additionally, the results indicated that EDI was mainly adopted because of cost and time savings as well as to reinforce the competitive position. 76.6% of the EDI adopters introduced EDI to meet the needs of their business partners.
90% of the companies fully automated forwarding of incoming EDI messages to concerning business applications. However, every second company reported manual interaction caused by problems with EDI processing.

Keywords: interoperability, DSML, UML, metamodeling, mappingmodel, WebML
Astract: Current software development projects comprise the development of complex software systems under immense time pressure. In the past decade, model-driven software development (MDSD) has become mainstream to tackle these challenges. In MDSD, domain specific modelling languages (DSML) are becoming more and more important. These languages allow to concisely represent all the peculiarities of a given domain in a model.
But being so specific, interoperability is needed with standardized modeling languages such as UML, because they offer a more common way of communication between different stakeholders. At the moment, interoperability can only be achieved by manually creating transformations between DSMLs and UML which is a challenging task.
This thesis presents a tool named ProfileGen, which tackles this challenge by proposing a semi-automatic approach for generating such transformations needed for interoperability between DSMLs and UML. In particular, a mapping language is presented which allows to manually link DSML elements with UML elements on a high-level of abstraction.
From such mappings, a generator framework automatically creates all artifacts needed for interoperability, including transformations from the DSMLs to UML and vice versa, as well as UML profiles for ensuring information loss free transformations .The approach is evaluated by a real-world case study, namely integrating WebML (a DSML for data-intensive Web applications) with UML.

Keywords: Refactoring, ATL, Model Transformations
Astract: The ATLAS Transformation Language (ATL) is a Language to describe rule-based mode-tomodel transformations. The more complex, different and interwoven the source and target model are, the more comprehensive the rules get. This results in detoriorated readability, maintainability and reusability and longer execution times of the transformations. At the moment transformations can only be modified by applying atomic operations. Combined operations have not yet been developed that would allow to improve transformations by restructuring.
The context of this thesis is the development of restructurings for ATL, so called refactorings, on the basis of existing work. These enable to transform ATL-transformations in a way that non-functional requirements and software quality attributes are improved.
A catalogue of refactorings was developed which separates into renaming, restructuring and optimizations for runtime speedup. There is a possibility to add domain-specific refactorings.
At the beginning existing transformations were analyzed and common 'bad smells' were derived by metrics and quality attributes like readability, maintainability and reusability. From this derivation the definition of the refactoring catalogue was created, which provides measures to eliminate bad smells. The catalogue is implemented using the ATL Refining Mode and is compared to an implementation in EMF Modeling Operations. The transformations of the ATLimplementation are metrically analyzed before and after each refactoring. Additionally the runtime of the transformations before and after the refactoring is timed and the number of executed bytecode statements is rated. The refactorings were evaluated regarding the application to other transformation languages (e.g. QVT-Relations).

Keywords:
Astract: Model-Driven Engineering promotes the use of models to conduct the different phases of the software development. In this way, models are transformed between different languages and notations until code is generated for the final application. Hence, the construction of correct Model-to-Model (M2M) transformations becomes a crucial aspect in this approach. Even though many languages and tools have been proposed to build and execute M2M transformations, there is scarce support to specify correctness requirements for such transformations in an implementation-independent way, i.e., irrespective of the actual transformation language used. In this paper we fill this gap by proposing a declarative language for the specification of visual contracts, enabling the verification of transformations defined with any transformation language. The verification is performed by compiling the contracts into QVT to detect disconformities of transformation results with respect to the contracts. As a proof of concept, we also report on a graphical modeling environment for the specification of contracts, and on its use for the verification of transformations in several case studies.

Keywords: