A Formalized Classification Schema for Model Consistency

Abstract

Although consistency is a widely used term, there is no common understanding in the modeling community on what consistency precisely means for model-based engineering. Nevertheless, most engineers would agree that models should be consistent to a certain degree to be useful. Although there are use case specific definitions of consistency and its properties, none of them are generally and independently applicable. As a result, we observed that engineers rarely employ a precise definition for their consistency notion. In addition, most engineers consider that models are consistent, neglecting that consistency usually relates model elements instead. To remedy these issues, we established a focus group of domain experts to derive a generally applicable and fine-grained notion of consistency. Using this fine-grained notion of consistency, we introduce a formalized classification schema for consistency covering seven dimensions. Afterwards this classification schema was applied by each domain expert to one of seven widely different model-based engineering scenarios. In turn, our schema provides a precise model-element-based definition of consistency and vocabulary for the modeling community to distinguish consistency relations. Above all, we aim to raise awareness for different consistency notions and facilitate a common understanding of consistency in the MBE community.

Download PDF

BibTeX

@techreport{kuehn_formalized_2023,
    author       = {Kühn, Thomas and Fuchß, Dominik and Corallo, Sophie and König, Lars and Burger, Erik and Keim, Jan and Mazkatli, Manar and Sağlam, Timur and Reiche, Frederik and Koziolek, Anne and Reussner, Ralf},
    year         = {2023},
    title        = {A Formalized Classification Schema for Model Consistency},
    doi          = {10.5445/IR/1000161127},
    institution  = {{Karlsruher Institut für Technologie (KIT)}},
    issn         = {2190-4782},
    series       = {Karlsruhe Reports in Informatics},
    pagetotal    = {37},
    language     = {english},
    volume       = {2023},
    number       = {1}
}