Für eine korrekte Darstellung dieser Seite benötigen Sie einen XHTML-standardkonformen Browser, der die Darstellung von CSS-Dateien zulässt.

. .

A delta-based modeling framework for SPLs


Model-driven software development (MDD) and Software Product Line Engineering (SPL-E) are widely used approaches to handle the complexity of software systems during development and to reduce time-to-market requirements. In model-driven development of software product lines models replace source code as primary executable artefacts in a reusable way. Variability has to be expressed in models when designing and implementing components of a SPL and bound when building a certain product.

The SiPL framework provides a difference-based delta-modeling approach exploiting recent advances in model differencing technologies in order to modeland bind variability.

Difference-based Delta-Modeling

Delta-Modeling is a general, language-independent, transformational approach to model and bind variability in the solution space of a SPL. The basic idea is to transform a model which represents a valid product of the configuration space into another model representing also an valid product. Therefore, features of the problem space, e.g. given by a feature model, are mapped onto delta-modules containing the respective transformation rules we referred to as edit steps. The mapping is realized by an delta-module application condition which is a propositional formula over a subset of all features, i.e. a delta-module can implement multiple features and a feature can be implemented by multiple delta-modules.

Usually, delta-modules must be manually written using a specific textual delta-language which can be considered as DSL for a specific modeling domain. In our approach a delta-module is created by modifying a given model and deriving the respective edit steps by comparing the two states of the model using the model versioning framework SiLift. As a result of the underlying graph transformation concepts we are able to analyze pairs of delta-modules for multiple relation kinds and exploit this relation analysis results for various operations on delta-modules for optimization of the delta-module set.