by Andreas Demuth
Abstract:
Model-Driven Engineering (MDE) is an established paradigm in software engineering. It promotes the use of models as primary development artifacts, thereby raising the level of abstraction and allowing stakeholders to focus on those parts of a system that seem most relevant to them without being overwhelmed by its high overall complexity. However, using models as abstractions of real-world systems does impose the necessity of using defined syntax and semantic. In order to achieve positive effects through the application of MDE, it is crucial that models do comply to those notations and are semantically valid. Unfortunately, ensuring that only valid models are built is a non-trivial task. Although various approaches have been developed to address this issue, those approaches are often limited to specific scenarios or restrict the model designer or the modeling process significantly. In this thesis, we investigate an approach that proposes the incremental generation and management of model syntactic and semantic constraints based on existing design models. We show how our approach can provide assistance to the designer for reaching a valid model without restraining the designer or the development process being used. The approach is applied to the domains of model-transformation, metamodeling, and software product lines to demonstrates both its feasibility and applicability. Case studies with industrial-size models suggest that the approach scales and that it is suitable for providing instant feedback about a model validity.
Reference:
Automatic Model-based Management of Design Constraints (Andreas Demuth), PhD thesis, Johannes Kepler University (JKU), Linz, Austria, 2013.
Bibtex Entry:
@PHDTHESIS{Demuth2013,
author = {Andreas Demuth},
title = {Automatic Model-based Management of Design Constraints},
school = {Johannes Kepler University (JKU), Linz, Austria},
year = {2013},
abstract = {Model-Driven Engineering (MDE) is an established paradigm in software
engineering. It promotes the use of models as primary development
artifacts, thereby raising the level of abstraction and allowing
stakeholders to focus on those parts of a system that seem most relevant
to them without being overwhelmed by its high overall complexity.
However, using models as abstractions of real-world systems does
impose the necessity of using defined syntax and semantic. In order
to achieve positive effects through the application of MDE, it is
crucial that models do comply to those notations and are semantically
valid. Unfortunately, ensuring that only valid models are built is
a non-trivial task. Although various approaches have been developed
to address this issue, those approaches are often limited to specific
scenarios or restrict the model designer or the modeling process
significantly. In this thesis, we investigate an approach that proposes
the incremental generation and management of model syntactic and
semantic constraints based on existing design models. We show how
our approach can provide assistance to the designer for reaching
a valid model without restraining the designer or the development
process being used. The approach is applied to the domains of model-transformation,
metamodeling, and software product lines to demonstrates both its
feasibility and applicability. Case studies with industrial-size
models suggest that the approach scales and that it is suitable for
providing instant feedback about a model validity.},
file = {:PhD Theses\\2013 Andreas Demuth\\dissertation_demuth-final_17092013.pdf:PDF},
keywords = {FWF P21321, FWF P23115},
owner = {AK117794},
timestamp = {2015.09.22}
}