Conclusion - Model-Driven and Software Product Line Engineering - page 244

obviously complicates the modeling and the derivation tool
chain.
The FieSta approach is our MD-SPL chain coping with
fine-grained variations and configurations. In this section, we
analyze the advantages and drawbacks of FieSta regarding the
MD-SPL engineering mechanisms. We focus on two aspects
impacted by fine-grained variability: i) variability expression
and product configuration in MD-SPL, and ii) the derivation of
configured products. We also compare FieSta with other MD-
SPL approaches.
9.2.1. Metamodeling and feature modeling
We use metamodeling and feature modeling for capturing
and expressing variability. Metamodels facilitate modeling
variations at the language level. Product designers, for
instance, building architects, are capable of configuring
differentproductsbycreatingdiversebuildingmodels.Feature
modeling allows us to configure products by selecting features.
Therefore, for instance, facilities designers and software
architects can configure products without the need for creating
complex models.
Using feature modeling and metamodeling separately gives
us the flexibility and power of expression of metamodels, and
thesimplicityoffeaturemodels.Wehavealsoproposedtorelate
metamodels and feature models to create what we have named
constraint models. Constraint models allow us to express fine-
grained variations between products of MD-SPL. We have
shown how to express the possible fine-grained variations
between products of an MD-SPL by creating relationships
between metamodels and feature models. For example,
fine-grained variations allow us to express that two Smart-
Home systems could be different in the location of their
automatic windows.