Christian Wende

FeatureMapper: mapping features to models

Variability modelling with feature models is one key technique for specifying the problem space of Software Product Lines (SPLs). To allow for the automatic derivation of a concrete product based on a given variant configuration, a mapping between features in the problem space and their realisations in the solution space is required. It is crucial to support the developer in the complex task of defining such mappings. These mappings can also be used to provide visualisations of the variant space that allow to reason over variability in SPLs. In this paper we present FeatureMapper, a tool that allows for defining mappings of features to model elements specifying feature realisations. These feature realisations can be defined in arbitrary Ecore-based languages. Furthermore, the tool supports different visualisation techniques that can help developers understand the complex designs of SPLs.

Derivation and Refinement of Textual Syntax for Models

Textual Syntax (TS) as a form of model representation has made its way to the Model-Driven Software Development community and is considered a viable alternative to graphical representations. To support the design and implementation of text editing facilities many concrete syntax and model mapping tools have emerged. Despite the maturity of these tools, users still spend considerable effort to specify syntaxes and generate editors even for simple metamodels. To reduce this effort, we propose to refine a specification that is automatically derived from a given metamodel. We argue that defaults in a customisable setting enable developers to quickly realise text-based editors for models. In particular in settings where metamodels evolve, such a procedure is beneficial. To evaluate this idea we present EMFText [1], an EMF/Eclipse integrated tool for agile Textual Syntax (TS) development. We show how default syntax can easily be tailored and refined to obtain a custom text editor for EMF models and demonstrate our approach by two examples.

Closing the gap between modelling and java

Model-Driven Software Development is based on standardised models that are refined, transformed and eventually translated into executable code using code generators. However, creating plain text from well-structured models creates a gap that implies several drawbacks: Developers cannot continue to use their model-based tool machinery, relations between model elements and code fragments are hard to track and there is no easy way to rebuild models from their respective code. This paper presents an approach to bridge this gap for the Java programming language. It defines a full metamodel and text syntax specification for Java, from which a parser and a printer are generated. Through this, Java code can be handled like any other model. The implementation is validated with large test sets, example applications are shown, and future directions of research are discussed.

A Framework for Generating Query Language Code from OCL Invariants

The semantical integrity of business data is of great importance for the implementation of business applications. Model-Driven Software Development (MDSD) allows for specifying the relevant domain concepts, their interrelations and their concise semantics using a plethora of modelling languages. Since model transformations enable an automatic mapping of platform independent models (PIMs) to platform specific models (PSMs) and code, it is reasonable to utilise them to derive data schemas and integrity rules for business applications. Most current approaches only focus on transforming structural descriptions of software systems while semantical specifications are neglected. However, to preserve also the semantical integrity rules we propose a Query Code Generation Framework that enables Model-Driven Integrity Engineering. This framework allows for mapping UML models to arbitrary data schemas and for mapping OCL invariants to sentences in corresponding declarative query languages, enforcing semantical data integrity on implementation level. This supersedes the manual translation of integrity constraints and, thus, decreases development costs while increasing software quality.

Reference attribute grammars for metamodel semantics

While current metamodelling languages are well-suited for the structural definition of abstract syntax and metamodelling platforms like the Eclipse Modelling Framework (EMF) provide various means for the specification of a textual or graphical concrete syntax, techniques for the specification of model semantics are not as matured. Therefore, we propose the application of reference attribute grammars (RAGs) to alleviate the lack of support for formal semantics specification in metamodelling. We contribute the conceptual foundations to integrate metamodelling languages and RAGs, and present JastEMF - a tool for the specification of EMF metamodel semantics using JastAdd RAGs. The presented approach is exemplified by an integrated metamodelling example. Its advantages, disadvantages and limitations are discussed and related to metamodelling, attribute grammars (AGs) and other approaches for metamodel semantics.

A role-based approach towards modular language engineering

Modularisation can reduce the effort in designing and maintaining language specifications. Existing approaches to language modularisation are typically either focused on language syntax or on language semantics. In this paper, we propose a modularisation approach covering both syntax and semantics. We propose defining composition rules on the level of abstract syntax, making it the central artefact in a language module. To enable clean interfaces for such language modules—effectively making them language components—we use role-modelling at the metamodel level. We discuss how role-based metamodelling supports the aspectual modularisation of language semantics and can also be integrated with concrete syntax specifications to build self-contained language components. We present the implementation of our approach in the LanGems language compositions system and show how it can be used to provide a modularised definition of the Object Constraint Language.

Validation of families of business processes

A Software Product Line (SPL) is a set of programs that are developed as a whole and share a set of common features. Product lines variability is typically specified using problem space models (i.e., feature models), solution space models that specify the realization of functionality and mapping models that link problem and solution space artifacts. In this paper, we consider this concept in the scope of families of business processes, whose specificity is that the solution space is defined with business process models. Solution space models are typically specified as model templates, and thus in the rest of the paper we will refer to business process model templates. While the previous research tackled the concepts of families of business processes, there have been very limited research on their validation.

Model-Based Language Engineering with EMFText

Model-based techniques are in wide-spread use for the design and implementation of domain specific languages (DSLs) and their tooling. The Eclipse Modeling Framework (EMF) is a frequently used environment for model-based language engineering. With its underlying modelling language Ecore, its XML serialisation support and its versatile extensibility it provides a solid grounding for many task-specific language development tools. In this tutorial, we give an introduction to model-based language engineering using EMFText, which allows users to develop powerful textual editors for Ecore-based DSLs that are tightly integrated with the EMF.

Extending variability for OCL interpretation

In recent years, OCL advanced from a language used to constrain UML models to a constraint language that is applied to various modelling languages. This includes Domain Specific Languages (DSLs) and meta-modelling languages like MOF or Ecore. Consequently, it is rather common to provide variability for OCL parsers to work with different modelling languages. A second variability dimension relates to the technical space that models are realised in. Current OCL interpreters do not support such variability as their implementation is typically bound to a specific technical space like Java, Ecore, or a specific model repository. In this paper we propose a generic adaptation architecture for OCL that hides models and model instances behind well-defined interfaces. We present how the implementation of such an architecture for DresdenOCL enables reuse of the same OCL interpreter for various technical spaces and evaluate our approach in three case studies.

DropsBox: the Dresden Open Software Toolbox

The Dresden Open Software Toolbox (DropsBox) is a software modelling toolbox consisting of a set of open source tools developed by the Software Technology Group at TU Dresden. The DropsBox is built on top of the Eclipse Platform and the Eclipse Modeling Framework. The DropsBox contributes to the development and application of domain-specific language changes (DSLs) in model-driven software development. It can be customised by tool and language developers to support various activities of a DSL’s life cycle ranging from language design to language application and evolution. In this paper, we provide an overview of the DSL life cycle, the DropsBox tools, and their interaction on a common example. Furthermore, we discuss our experiences in developing and integrating tools for DropsBox in an academic environment.