Further Reading - Model-Driven and Software Product Line Engineering - page 225

selected target technological platform. The execution ordering
of the model transformations is predefined by creating a
type of abstract execution ordering. The abstract execution
ordering defines the required sequence of calls to abstract
transformation rules. The concrete transformation rules are
executed once the groups of model transformations to be used
are defined from the selected variants and the selected target
technological platform.
To replace the abstract transformation rules by the
concrete transformation rules during execution of the model
transformations, the authors propose a composition technique
they call module superimposition . To apply this technique,
transformation rules must be grouped in modules . This
technique allows the modification of an execution ordering,
which includes transformation rules from a module “m-1”,
overriding it to include: (1) new calls to transformation
rules from a module different from “m-1”, and (2) calls
to transformation rules with the same names and the
same parameters that are included in the module “m-1”,
but from a module different from “m-1”. This mechanism
has been implemented using the ATLAS Transformation
Language (ATL) [JOU 05].
8.9.3. Loughran et al.'s approach
In [LOU 08, SAN 08] the authors propose a powerful
approach that relies on some feature-oriented models and
programming.
Problem space: expressing variability and configuring
products . Loughran et al . propose an approach where
variabilityisexpressedusingcardinality-basedfeaturemodels.
Products are configured creating feature configurations.
The main purpose of Loughran et al . is to provide support
for composition of software components based on feature