Keywords:
Astract: Complex service-oriented systems typically span interactions between people and services. Compositions in such systems demand for flexible interaction models. In this work we introduce an approach for discovering experts based on their dynamically changing skills and interests. We discuss human provided services and an approach for managing user preferences and network structures. Experts offer their skills and capabilities as human provided services that can be requested on demand. Our main contributions center around an expert discovery method based on the concept of hubs and authorities in Web-based environments. The presented discovery and interaction approach takes trust-relations and link properties in social networks into account to estimate the hub-expertise of users. Furthermore, we show how our approach supports flexible interactions in mixed service-oriented systems.

Keywords: model-driven, model versioning, conflict detection, refactorings
Astract: Within the last years, the popularity and appearance of model-driven software development (MDSD) has increased significantly.
The model-driven paradigm induces a shift from code-centric to model-centric development. Software models are considered as first class entites building the basis for generating an executable software.
Nowadays, MDSD is a widely accepted technology helping software developers to accelerate and simplify their work by raising the abstraction level of the software artefacts they create.
Large-scale software is usually produced by a high number of software developers, who are often working in parallel and spread all over the world. Therefore, the version management of software models is crucial for an effective, collaborative software development. One of the most important components of a versioning system for model artefacts is a precise difference and conflict detection, which is capable of identifying all operations executed by the developers and the eventually resulting conflicts that arise at merging independently changed models.
Without an appropriate difference and conflict detection system, software developers are forced to find occuring problems manually, which is a time consuming process. Especially complex operations, refactorings and language specific conflicts are hard to identify and can lead to serious malfunctions in the produced software if they remain undiscovered.
The main goal of this work is the development and implementation of an intelligent and adaptable conflict detection mechanism for models in the context of versioning. A framework will be presented which is able to precisly calculate the differences and detect conflicts in a generic way. Users may extend this system by providing language specific definitions which will be used to refine the functionality for specific modeling languages. In that way, the quality of the difference and conflict detection can be improved to minimize the need for manual interaction.

Keywords: QVT, Model, Model Transformation, Debugging, Visualization
Astract: Model transformations (MT) play a key role in the Model Driven Engineering (MDE) paradigm, leading to the standardization of the Query/View/Transformation (QVT) model transformation language by the Object Management Group (OMG). Until now, however, this language did not attract the same interest as the Unified Modeling Language (UML), because of the lack of adequate debugging facilities which are necessary regarding the following three problem areas: First, declarative languages like QVT Relations (QVT-R) hides the operational semantics of transformations. Only the information provided by the interpreter, as well as the tendered inputs and returned outputs are available for tracking the progress of transformations. Furthermore, the ordering of transformation application is hidden by the MT engines providing only a black-boxes view to the users. This can lead to the problem of impedance mismatches between design and runtime. These characteristics of QVT-R are assets for developing, but are handicaps for debugging. Second, QVT-R code is specified on higher abstraction level than its execution and state-of-the-art debugging. This deteriorates the ability to deduce causes from produced results. Third, the information content responsible for operating MTs is spread over several artifacts including the input model, a resulting target model and the QVT-R code. As a consequence, the reasons for a particular outcome are hard to be derived from the involved artifacts. This severely harms the ease of debugging.
Therefore, this master thesis tackles the mentioned problems by visualizing QVT-R as Transformations Nets, using the MT framework ``Transformations On Petri Nets In Color'' (TROPIC) based on Colored Petri Nets (CPN). This can be seen as explicit definition of operational semantics on a high abstraction level providing a white-box view for debugging QVT-R. This thesis proposes a procedure model formulated in a conceptual approach and in a prototypic implementation striving for bridging the existing gap between these two different paradigms by mapping the concepts of QVT Relations to such nets. In this thesis three particular contributions are provided: (i) a solution approach for unidirectional mappings producing target models from an existing source model, (ii) the support for model inheritance, (iii) and synchronization approaches for timely and version-based incremental changes.

Keywords: Mashups, Semantic Web, SOA, BPEL, XForms
Astract: Managing lifetime memories has been a blurry challenge ever since it was originally proposed by Vannevar Bush in his seminal paper (1945). At the time, the concept was extremely speculative and indeed, Memories for Life has remained a challenging problem for the next five decades. Recently this idea was declared one of the grand challenges in computing research where its scope has been expanded and the problem has been outlined as a multidisciplinary problem that should address issues such as knowledge and databases, information retrieval, security/privacy, and human-computer interaction. New possibilities for Memories for Life emerged with the introduction of Semantic Web technologies that make it possible to model a diverse range of memory items and their associations in so-called Semantic Desktops. Today, there are a handful of Semantic Desktops that are to some extent able to organize and manage personal life items. However, their application is limited to the storage and retrieval of various personal items. The next step towards the goals of the Memories for Life grand challenge is to integrate this organized information in the business and social network services. With the emergence of complex business processes, there is a growing need to automatically map and embed the user's information context into global business services. This aspect of lifetime memories, which may have a great effect on extending the footprint of such memories for interactions with global and social network services, has not yet been explored. In order to realize such use cases, the major obstacle to face, is providing a secure and comprehensible service layer that can be easily used by both end-users and business processes. In this proposal some information integration approaches will be introduced and among them the Mashups architecture will be explored in more details. The Mashup architecture utilizes the advantages of Web 2.0 that has recently hit the mainstream of Internet computing and continues its rapid evolution. Web 2.0 has introduced new possibilities for a better human computer interaction via rich applications such as Mashups that provide a user-driven micro-integration of web-accessible data. Mashup envisions building effective and light-weight information processing solutions based on the exposed Web Services of organizations. Such Web Services may range from simple Web Services such as RSS and REST based services to complex BPEL services for more serious use cases.
The ultimate goal of this dissertation is to bridge the gap between the personal information world and the global business world by introducing semantically-aware Mashups that on one hand facilitates the secure and intuitive creation and sharing of information and new services based on lifetime personal information, and on the other hand fosters the Internet computing paradigm via business processes that take advantage of user-generated personal services.

Keywords: knowledge management, Web applications, CAN data, MSSQL Server, ASP.NET, patterns, layered architecture, LINQ, object-relational mapping, schema evolution
Astract: A sustainable knowledge management and sophisticated tool support are extremely important for every successfully operating company. Within an extremely wide range of tools, Web applications gain more and more importance to meet the constantly increasing requirements.
Outstanding benefits are worldwide accessibility and interoperability on a very large scale. Moreover, in many cases no installation on single clients is needed, updates can be provided easily, and centralized data management on the server-side avoids costly synchronizations. These factors lead to reduced costs for the information infrastructure and support the employees to do their work. For example, a constructor of fire-fighting trucks can query the headquarter's database on the other side of the world to find out which CAN data (Controller Area Network) is needed to configure a specific vehicle.
Building such solutions is not a trivial task and therefore disciplines like Web Engineering and Internet Computing emerged. Furthermore, developers can choose from a wide range of technologies to realize their solutions. Handling these technologies leads to successful development of Web applications. This master's thesis describes in detail the solution for a specific problem in the industry, namely a Web application called CORA (CAN Bus Organization - Rosenbauer Assistant) to manage the CAN data for Rosenbauer International AG. According to the company's IT-infrastructure MSSQL Server, IIS, and ASP.NET were chosen as core technologies. On one hand the .NET framework provides the possibility to develop applications in a rather short time, also known as Rapid Application Development (RAD). On the other hand many RAD-techniques are not applicable on large enterprise solutions where the complexity has to be broken down into several layers. This work presents approaches, patterns and techniques for each layer.
The Data Access Layer is responsible for retrieving and manipulating data and uses LINQ to SQL as object-relational mapper. The Business Layer handles the communication to external systems and adds business functionality between the Data Access Layer and the Presentation Layer.
Finally, the Presentation Layer presents the data in an appropriate format and handles the user interaction.
Furthermore, the database schema of the sample application has been constantly renewed and improved to cover additional requirements like multiple CAN bus systems, multilingualism, user administration, and a history of important entities. Therefore, schema evolution and data migration play an additional important role in this thesis. All these aspects are elaborated theoretically and explained practically with the help of CORA.

Keywords: Anonymization, Pseudonymization, Research and Teaching, l-Diversity, Diversity, k-Anonymity, Data Privacy, Security
Astract: Personal Data, concerning unique individuals, requires accurate treatment ever since. To avoid the disclosure of these unique individuals through the exploit of personal data during electronic ex- change, certain steps can be taken, including anonymization- and pseudonymization-techniques. In the first instance we focus on the comparison of current anonymization- and pseudonymization- techniques and then continue with the proposal for a verifying-component which is going to be rea- lized within the scope of a case-study at the Technical University of Vienna. Purpose of the compo- nent is the delivery of a decision-base for the peer-to-peer exchange of medical data so no unique individuals can be identified during the exploit of this data. Therefore the component assures a cer- tain standard of pseudonymity for the medical data in regards of diversity and anonymity. Finally a proposal for the verifying-component is made in the last section of this scientific work. Re- search showed that current anonymization and pseudonymization methods are insufficient in terms of security and attacks on them, so that a maximum level of security cannot be guaranteed at this point of time.

Keywords: accessibility, web, software engineering, code generation
Astract: Model engineering is a current trend in software engineering.
Its aim is the automatic generation of executable code from abstract software models. Latest techniques focus mainly on functional aspects of the generated code, e.g. how Java code is generated from UML diagrams.
However, aspects of quality which do not come from the technical requirements are rarely considered. An important aspect of quality represents among other things the accessibility of web pages.
Legislation in Austria requires a barrier free access to official internet appearances for humans with special needs. This should have been implemented until 1st of Jan. 2008. In order to determine the accessibility of web applications so called Web Accessibility Guidelines (WAG) have been developed. For example, there are the WCAG (Web Content Accessibility Guidelines)of the W3C, which already became generally accepted as a de facto web accessibility standard. In order to fulfil the conformity with the legal regulations demanded above, public internet appearances must correspond to the respectively demanded guidelines. In this diploma thesis it is evaluated, in which way modern component frameworks and code generators fulfil the requirements resulting from the given guidelines. Therefore, WAG conform static html pages are used as a starting point. These static pages are developed to dynamic web applications by means of chosen component frameworks and code generators. The result of the generation process of each chosen development technology is finally rendered in form of html pages in the browser. These rendered pages are evaluated by means of WAG evaluation tools (test programs for the WAG conformance) regarding guidelines selected before. The results of this conformance evaluation demonstrate, to which extent the selected Web accessibility guidelines are violated. This evaluation also highlights which of the used development environments are better in terms of WAG conformance.

Keywords: RIA, e-Learning, UML, Visualization, Diagrams, Class Diagrams, Modeling, Metamodeling, Model Engineering, Web
Astract: The ability of software engineers to develop clear and comprehensive models according to a specific problem domain has become an essential factor for the success of software projects. The trend of Model Driven Software Engineering (MDSE) even increases this factor.
Universities have to qualitatively educate their students in this field to prepare them for those challenges.
At the Vienna university of Technology (VUT), the course "Object Oriented Modeling" teaches the basics of modeling and the Unified Modeling Language (UML). Due to the high number of students, various e-Learning elements have already been established, mainly for teaching the theory behind UML.
The goal of this thesis was to develop a web-based modeling tool, that is integrated into the e-Learning platform of the VUT using Rich Internet Application technologies. Thus, students can solve practical modeling exercises comfortably and directly using the e-Learning platform. Therefore, a generic approach shall be used by utilizing metamodeling: Visualization and manipulation of two-dimensional diagrams is based on certain patterns. A framework was created, that allows for creating complete diagram editors with drag-and-drop functionality, by specifying the structure of the diagram elements (abstract syntax) and the visual notation elements (concrete syntax). This enables to easily create various UML editors and to align the notation elements with those of the course. Within this thesis, three editors (class diagram, state diagram, and sequence diagram) have been created with the aid of the framework and integrated into the e-Learning platform.

Keywords: business documents, XML Schema, model-driven development, UN/CEFACT's Core Components, UML Profile for Core Components, UN/CEFACT's Modeling Methodology, service-oriented architectures, process choreographies, domain specific languages, UML
Astract: Automated business-to-business (B2B) interactions between companies are constantly superseding old paper-based processes. This automation of inter-organizational processes requires a two-fold agreement between the participating business partners. First of all, business partners must agree on a common process choreography, unambiguously defining the exact exchange order of business documents in an inter-organizational business process. Consequently, business partners must agree on the structure of the exchanged business information as well. The two main business document paradigms, which are we elaborate on in this thesis, are top-down business document standards and bottom-up business document standards. The research question, this thesis aims to solve, is how to provide appropriate methods for the definition of business documents for inter-organizational business processes. Due to their special characteristics, such as the involvement of various stakeholders from different companies, the definition of business documents for inter-organizational business processes is not as straightforward as the definition of business documents for intra-organizational business processes. For the definition of inter-organizational business documents we employ two different approaches - a top-down approach and a bottom-up approach. For both approaches we provide appropriate methods for the definition of business documents, the mapping between different business document definitions, and the derivation of XML-based deployment artifacts from business document definitions. We thoroughly cover state-of-the art in the domain of inter-organizational business documents and inter-organizational business processes. This thesis starts by giving an introduction to the domain of Electronic Data Interchange (EDI) and shows the transition of data-centric EDI solutions to modern B2B systems in Chapter 1. Thereby, the specific requirements for B2B interactions are elaborated - in particular in regard to the definition of the exchanged business information. We motivate the findings of this thesis using an accompanying example from the domain of pan-European waste transport, which is introduced in Chapter 2. In Chapter 3 we provide a survey of current state-of-the-art in business documents standards. Chapter 4 provides an introduction to UN/CEFACT's Core Components, a top-down business document standard which is key to this thesis. In a nutshell, core components are implementation neutral building blocks for assembling business documents. Although this implementation neutrality is one of the strengths of core components, it hinders a broad diffusion of the standard, since no common representation format for core components is available. To address this issue, this thesis provides three reference representation formats for core components: (i) a UML Profile for Core Components (Chapter 5), (ii) a Domain Specific Language for Core Components (Chapter 6), (iii) and a Web Ontology Language representation for Core Components (Chapter 7). The derivation of XML Schema artifacts from core components, which may be deployed to IT systems, is covered in Chapter 8. A successful diffusion of core Components may only be achieved if a broad user community has access to the necessary core component definitions. Consequently, we provide a registry meta-model for core components in Chapter 9. Bottom-up business document standards are subject to discussion in the second part of this thesis. Thereby, domain specific extensions to bottom-up document standards are introduced in Chapter 10. Consequently, we examine how to map bottom-up standard definitions to core component based top-down standard definitions in Chapter 11. Finally, we show how to combine business choreography models and business document models in Chapter 12.
Related work is discussed in Chapter 13 and Chapter 14 concludes the contributions of this thesis. In summary this thesis provides the following seven contributions: (1) An overview of business document standards, based on standard clusters; (2) three reference representation formats for core components using the Unified Modeling Language, Domain Specific Languages, and Web Ontology Language for Core Components; (3) an unambiguous derivation of XML Schema artifacts from UML based core component models; (4) a registry meta-model for a core component registry; (5) domain specific extensions for an XML based bottom-up business document standard; (6) a mapping of bottom-up business document standards to top-down business documents standards; (7) an integration of UML based core component models in UML based business choreography models. In short the overall approach facilitates the definition of business documents in an inter-organizational context and fosters reuse of existing business document definitions.

Keywords: Electronic Commerce, B2B, Business Process Modeling, Business Modeling, UML, Model-Driven Software Development, Business-to-Business
Astract: Business-to-business (B2B) electronic commerce builds upon inter-organizational business processes that cross the borders of enterprises. Their design and implementation presupposes a different approach than intra-organizational processes do. Experience shows that bottom-up approaches starting from the IT layer of a single enterprise - expecting that all other business partners adjust to it - do not work out. Instead, a prolific B2B design approach must consider three layers in a top-down manner: First, the economic perspective identifies the players and their value exchanges within a business network resulting in a business model. Second, business collaboration models specify the choreography of inter-organizational business processes in accordance with the business model. Finally, the business collaboration models are transformed to deployment artifacts to be interpreted by IT systems. In this thesis, we propose a design approach for B2B integration based on the Unified Modeling Language (UML) considering all three layers.