Ralph Depke

Process-Oriented, Flexible Composition of Web Services with UML

The composition of elementary web services to larger-scale services has become an important means to enhance e-business collaborations. If such composite web services can also integrate legacy components that are not yet provided as web services, the number of possible compositions is increased. Following a process-oriented approach, the compositions can be described as control- and data-flow between available web services and components. This paper discusses the Business Process Execution Language for Web Services (BPEL4WS), an existing service composition language, and proposes UML-WSC as an alternative, visual language. For the advanced description of service interfaces, UML-WSC extends the type system of the established Web Service Definition Language (WSDL).

Improving the agent-oriented modeling process by roles

The agent-oriented modeling process is divided in a typical sequence of activities, i.e., \emph{requirements specification}, \emph{analysis}, and \emph{design}. The \emph{requirements} are specified by descriptions of the systems functionality and by exemplary scenarios of essential interactions. In \emph{analysis} the systems structure is captured and mandatory behavior of agents is prescribed. The \emph{design} model describes system behavior by means of local operations. The problem arises how the transition between these different stages of the modeling process can be performed. In this paper, we introduce a concept of roles in order to support the transition in a systematic way and thereby improving the agent-oriented modeling process.

ROLES IN AGENT-ORIENTED MODELING

For the generic specification of protocols, goals, or workflows, many approaches to agent-oriented modeling provide a concept of role. Roles abstract from the concrete agents involved in an interaction. They provide means for the evolution of agents and serve as components of agent design. Despite the widespread usage of roles in agent-oriented modeling, a systematic analysis of the different aspects and properties of this concept is still missing. In this paper, we perform such an analysis and identify requirements for a general role concept. We develop such a role concept for a modeling approach based on the UML and graph transformation systems and exemplify its use for the specification (and application) of protocols. Finally, we provide a run-time semantics for roles based on concepts from the theory of graph transformation.

Agent-oriented modeling with graph transformation

The agent paradigm can be seen as an extension of the notion of (active) objects by concepts like autonomy, cooperation, and goal-oriented behavior. Mainstream object-oriented modeling techniques do not account for these agentspecific aspects. Therefore, dedicated techniques for agent-oriented modeling are required which are based on the concepts and notations of object-oriented modeling and extend these in order to support agent-specific concepts. In this paper, an agent-oriented modeling technique is introduced which is based on UML notation. Graph transformation is used both on the level of modeling in order to capture agent-specific aspects and as the underlying formal semantics of the approach.

Process-oriented, consistent integration of software components

The integration of software components becomes a more and more important issue in software engineering. Process-oriented approaches should provide automated information processes. Therefore, the software components have to be integrated in a consistent way, i.e., their export interfaces have to be respected by the importing components. Furthermore, the type system of component interfaces has to support a tunable degree of freedom. This allows the insertion of components with interfaces of restricted but sufficient degree of compatibility. In this paper, we develop a concept for consistent and flexible integration of components. We present a process modeling language that combines UML and XML in order to support consistent, flexible, and executable processes. Finally, we provide a formalization of the proposed component type system.

Ein Vorgehensmodell für die Multimedia-Entwicklung mit Autorensystemen

Zusammenfassung. Multimedia-Anwendungen sind interaktive Softwaresysteme und verlangen als solche, mit softwaretechnischen Methoden erstellt zu werden. Sie werden heutzutage in der Regel mit Hilfe von Autorensystemen entwickelt, die eine Ad-hoc-Entwicklung auf Implementierungsniveau unterstützen. Hierdurch und wegen des Fehlens etablierter Vorgehensmodelle für die Multimedia-Softwareentwicklung reduziert sich der Multimedia-Entwicklungsprozeß auf die Implementierungsphase. Dies führt zu den in der Softwaretechnik bekannten Problemen wie mangelnder Konzeption und fehlender Dokumentation. Wir stellen in diesem Beitrag ein Vorgehensmodell für die Entwicklung von Multimedia-Anwendungen vor, in dessen Mittelpunkt eine Analyse- und Entwurfsphase im Hinblick auf eine Implementierung der Multimedia-Anwendung mit einem Autorensystem stehen. Ausgehend von einem frameworkbasierten Analysemodell der Anwendung und einem Modell der Realisierungsmöglichkeiten mit einem konkreten Autorensystem wird systematisch ein Implementierungsmodell auf Instanzebene abgeleitet, das als Eingabe für das Autorensystem verwendet wird. Das postulierte Vorgehensmodell wird exemplarisch für das Autorensystem Director am Beispiel der Domäne multimedialer Lehr-/Lernanwendungen erläutert.Abstract. Multimedia applications are interactive software systems and should therefore be developed using software engineering techniques. They are nowadays built using authoring systems that support ad hoc implementation without a preceding modeling phase. Because of that and due to the lack of established process models for multimedia software development, the development process is reduced mainly to the implementation phase. This leads to problems well-known in software engineering like missing conceptualization and missing documentation. In this paper, we present a process model for the development of multimedia applications which is centered upon an analysis and a design phase with regard to an implementation with an authoring system. Starting from a framework-based analysis model of the application and a model of the programming capabilities of a concrete authoring system, an instance-level implementation model can be systematically derived. The proposed process model is exemplarily explained using the Director authoring system for the domain of computer-based training applications.

Modelling and Analysis of Agents’ Goal-Driven Behavior Using Graph Transformation

Extending mainstream object-oriented concepts, the agent paradigm promotes the concept of goals realized by means of strategies. To account for such specific aspects, dedicated techniques for agent-oriented modelling are required which go beyond standard techniques of object-oriented modelling.

Design of an agent-oriented modeling language based on graph transformation

The use of UML extension mechanisms for the definition of an Agent-Oriented Modeling Language only fixes its syntax. But agent concepts demand an appropriate semantics for a visual modeling language. Graphs have been shown to constitute a precise and general semantic domain for visual modeling languages. The question is how agent concepts can be systematically represented in the semantic domain and further on be expressed by appropriate UML diagrams. We propose a language architecture based on the semantic domain of graphs and elements of the concrete syntax of UML. We use the proposed language architecture to define parts of an agent-oriented modeling language.