Logo des Repositoriums

Auflistung nach Autor:in "Gerber, Stefan"

1 - 2 von 2
  • Konferenzbeitrag
    EA Model as central part of the transformation into a more flexible and powerful organisation
    (Enterprise modelling and information systems architectures – concepts and applications, 2007) Gerber, Stefan; Meyer, Uwe; Richert, Claus
    This report introduces an approach how Enterprise Architecture (EA) design can be deployed in a large financial organisation for strategic transformation. Our EA design embraces all main components of the business organisations, its information systems and the way they work to achieve business objectives. In order to tackle such EA design and its deployment, governance, design and measurement principles are required to keep EA consistent and avoid misunderstandings among stakeholders. Since EA focuses on a holistic view of the organisation, full EA deployment is risky due to cost and organisational impact. Therefore we use an iterative approach within EA deployment that will be considered as an assessment process evaluating the whole IT-landscape of a certain CIO area. There are metrics used which allow the identification of transformation objects and these will be reworked in different structures by using architectural principles and then integrated into EA. Finally the existing EA will be evaluated (together with transformation object) by EA design principles and either the transformation will be rejected or design principles will be adopted. In order to make this model operative it is embedded in an architecture organizational structure which is independent from the organizational structure of the enterprise.
  • Konferenzbeitrag
    Model Extraction: Transformation von Benutzeranforderungen in UMLModelle mit Hilfe statistisch-linguistischer Methoden und Heuristiken
    (INFORMATIK 2006 – Informatik für Menschen – Band 2, Beiträge der 36. Jahrestagung der Gesellschaft für Informatik e.V. (GI), 2006) Mai, André; Gerber, Stefan
    Benutzeranforderungen werden von den Benutzern häufig natürlichsprachlich formuliert, obwohl die UML als Quasistandard in Analyse- und Design von Softwareanwendungen gilt. Durch den unvermeidlichen Medienbruch sind Missverständnisse vorprogrammiert, die durch eine Transformation von natürlichsprachlichen Texten in semiformale Modelle vermieden werden können. Im Beitrag werden existierende Möglichkeiten zur Extraktion von UML-Klassenmodellen aus Benutzeranforderungen, statistisch-linguistische Analysen und Heuristiken untersucht und zu einer neuen Methode kombiniert.