Information Technology Reference
In-Depth Information
non-functional requirements. However, they are different in
manyways.Metamodelsfacilitatethemodelingofvariationsat
language level. Product designers who are domain experts are
capable of configuring different products by creating diverse
and rich application domain models. Thus, metamodeling
implies a constructive approach that requires a high level of
expertise.Still,modelsandmetamodelsaregeneralvehiclesfor
expressing variability in software artifacts, which are models
in MDE. On the other hand, feature models ease the modeling
of variations using a well-known syntax and semantics to
capture variability. Feature modeling allows the configuration
ofproductsonlybyselectingfeaturesandhidingthecomplexity
of building complex models; this is a selection-based approach
that requires only domain knowledge.
The use of alternative (less-used) mechanisms for capturing
and expressing variability, such as ontology models from the
SPLE field, seems very relevant. However, there are not
enough practical examples available to incorporate in the
MDE field. Therefore, an MD-SPL approach must provide at
leastmechanismstocapturevariabilityandconfigureproducts
based on meta modeling and feature modeling. In addition,the
approach must allow product line architects to decide when to
choose either metamodeling or feature modeling.
4.2.3.
Relating several points of view
We introduced the need for MD-SPL approaches for
providing mechanisms to represent, using metamodels and/or
feature models, each product line member according to
different concerns. Thus, a product line architect could, for
example, represent the variability of a Smart-Home system
by using a metamodel to capture the conceptual elements
of the application domain, and by using a feature model
to represent the scope of the product line for architectural
features. Consequently, a product designer could configure a
product by creating a model that conforms to the conceptual
Search WWH ::
Custom Search