Logo des Repositoriums

Auflistung nach Autor:in "Kramer, Max E."

1 - 3 von 3
  • Konferenzbeitrag
    Automatically binding variables of invariants to violating elements in an OCL-aligned xbase-language
    (Modellierung 2016, 2016) Fiss, Sebastian; Kramer, Max E.; Langhammer, Michael
    Constraints that have to hold for all models of a modeling language are often specified as invariants using the Object Constraint Language (OCL). If violations of such invariants shall be documented or resolved in a software system, the exact model elements that violate these conditions have to be computed. OCL validation engines provide, however, only a single context element at which a check for a violated invariant originated.Therefore, the computation of elements that caused an invariant violation is often specified in addition to the invariant declaration with redundant information. These redundancies can make it hard to develop and maintain systems that document or resolve invariant violations. In this paper, we present an automated approach and tool for declaring and binding parameters of invariants to violating elements based on boolean invariant expressions that are similar to OCL invariants. The tool computes a transformed invariant that returns violating elements for each iterator variable of the invariant expression that matches an explicitly declared invariant parameter. The approach can be used for OCL invariants and all models of languages conforming to the Meta- Object Facility (MOF) standard. We have evaluated our invariant language and transformation tool by transforming 88 invariants of the Unified Modeling Language (UML).
  • Zeitschriftenartikel
    Challenges to Trading-Off Performance and Privacy of Component-Based Systems
    (Softwaretechnik-Trends Band 36, Heft 4, 2016) Seifermann, Stephan; Yurchenko, Kateryna; Kramer, Max E.
    Determining privacy properties of software systems is essential for certification in certain domains and gains importance for users of software services. Late discovery of degraded privacy properties during development phases makes fixing issues hard and expensive. Approaches that focus on architectural privacy predictions are rare and often do not integrate well with existing tools for performance predictions so that trade-off analyses are not supported. In this paper, we suggest extending the Palladio Component Model (PCM) by means of modeling privacy requirements to support privacy predictions, code generation, and verification, as well as trade-off decisions. The goal of this integration with PCM is to support the development of quality-aware component-based architectures: Our approach will allow trade-offs for privacy and performance properties during the early design and will ease the verification of the implementation.
  • Konferenzbeitrag
    Synchronizing heterogeneous models in a view-centric engineering approach
    (Software Engineering 2014, 2014) Kramer, Max E.