Gregor Engels

Software Engineering for Self-Adaptive Systems: A Second Research Roadmap

The goal of this roadmap paper is to summarize the state-of-the-art and identify research challenges when developing, deploying and managing self-adaptive software systems. Instead of dealing with a wide range of topics associated with the field, we focus on four essential topics of self-adaptation: design space for self-adaptive solutions, software engineering processes for self-adaptive systems, from centralized to decentralized control, and practical run-time verification & validation for self-adaptive systems. For each topic, we present an overview, suggest future directions, and focus on selected challenges. This paper complements and extends a previous roadmap on software engineering for self-adaptive systems published in 2009 covering a different set of topics, and reflecting in part on the previous paper. This roadmap is one of the many results of the Dagstuhl Seminar 10431 on Software Engineering for Self-Adaptive Systems, which took place in October 2010.

Graph transformation for specification and programming

The framework of graph transformation combines the potentials and advantages of both, graphs and rules, to a single computational paradigm. In this paper we present some recent developments in applying graph transformation as a rule-based framework for the specification and development of systems, languages, and tools. After reviewing the basic features of graph transformation, we discuss a selection of applications, including the evaluation of functional expressions, the specification of an interactive graphical tool, an example specification for abstract data types, and the definition of a visual database query language. The case studies indicate the need for suitable structuring principles which are independent of a particular graph transformation approach. To this end, we present the concept of a transformation unit, which allows systematic and structured specification and programming based on graph transformation.

Dynamic meta modeling: a graphical approach to the operational semantics of behavioral diagrams in UML

In this paper, dynamic meta modeling is proposed as a new approach to the operational semantics of behavioral UML diagrams. The dynamic meta model extends the well-known static meta model by a specification of the systems dynamics by means of collaboration diagrams. In this way, it is possible to define the behavior of UML diagrams within UML. The conceptual idea is inherited from Plotkins structured operational semantics (SOS) paradigm, a style of semantics specification for concurrent programming languages and process calculi: Collaboration diagrams are used as deduction rules to specify a goal-oriented interpreter for the language. The approach is exemplified using a fragment of UML state-chart and object diagrams. Formally, collaboration diagrams are interpreted as graph transformation rules. In this way, dynamic UML semantics can be both mathematically rigorous so as to enable formal specifications and proofs and, due to the use of UML notation, understandable without prior knowledge of heavy mathematic machinery. Thus, it can be used as a reference by tool developers, teachers, and advanced users.

Detecting and Resolving Process Model Differences in the Absence of a Change Log

Business-driven development favors the construction of process models at different abstraction levels and by different people. As a consequence, there is a demand for consolidating different versions of process models by detecting and resolving differences. Existing approaches rely on the existence of a change log which logs the changes when changing a process model. However, in several scenarios such a change log does not exist and differences must be identified by comparing process models before and after changes have been made. In this paper, we present our approach to detecting and resolving differences between process models, in the absence of a change log. It is based on computing differences and deriving change operations for resolving differences, thereby providing a foundation for variant and version management in these cases.

A methodology for specifying and analyzing consistency of object-oriented behavioral models

Object-oriented modeling favors the modeling of object behavior from different viewpoints and the successive refinement of behavioral models in the development process. This gives rise to consistency problems of behavioral models. The absence of a formal semantics for UML models and the numerous possibilities of employing behavioral models within the development process lead to the rise of a number of different consistency notions. In this paper, we discuss the issue of consistency of behavioral models in the UML and present a general methodology how consistency problems can be dealt with. According to the methodology, those aspects of the models relevant to the consistency are mapped to a semantic domain in which precise consistency tests can be formulated. The choice of the semantic domain and the definition of consistency conditions can be used to construct different consistency notions. We show the applicability of our methodology by giving an example of a concrete consistency problem of concurrent object-oriented models.

Conceptual modelling of database applications using an extended ER model

Abstract In this paper, we motivate and present a data model for conceptual design of structural and behavioural aspects of databases. We follow an object centered design paradigm in the spirit of semantic data models. The specification of structural aspects is divided into modelling of object structures and modelling of data types used for describing object properties. The specification of object structures is based on an E xtended E ntity- R elationship (EER) model. The specification of behavioural aspects is divided into the modelling of admissible database state evolutions by means of temporal integrity constraints and the formulation of database (trans)actions. The central link for integrating these design components is a descriptive logic-based query language for the EER model. The logic part of this language is the basis for static constraints and descriptive action specifications by means of pre- and postconditions. A temporal extension of this logic is the specification language for temporal integrity constraints. We emphasize that the various aspects of a database application are specified using several appropriate, but yet compatible formalisms, which are integrated by a unifying common semantic.

Rule-Based Specification of Behavioral Consistency Based on the UML Meta-model

Object-oriented modeling favors the modeling of object behavior from different viewpoints and at different levels of abstraction. This gives rise to consistency problems between overlapping or semantically related submodels. The absence of a formal semantics for the UML and the numerous ways of employing the language within the development process lead to a number of different consistency notions. Therefore, general meta-level techniques are required for specifying, analyzing, and communicating consistency constraints. In this paper, we discuss the issue of consistency of behavioral models in the UML and present techniques for specifying and analyzing consistency. Using meta-model rules we transform elements of UML models into a semantic domain. Then, consistency constraints can by specified and validated using the language and the tools of the semantic domain. This general methodology is exemplified by the problem of protocol statechart inheritance.

ClassSheets: automatic generation of spreadsheet applications from object-oriented specifications

Spreadsheets are widely used in all kinds of business applications. Numerous studies have shown that they contain many errors that sometimes have dramatic impacts. One reason for this situation is the low-level, cell-oriented development process of spreadsheets.We improve this process by introducing and formalizing a higher-level object-oriented model termed ClassSheet. While still following the tabular look-and feel of spreadsheets, ClassSheets allow the developer to express explicitly business object structures within a spreadsheet, which is achieved by integrating concepts from the UML (Unified Modeling Language). A stepwise automatic transformation process generates a spreadsheet application that is consistent with the ClassSheet model. Thus, by deploying the formal underpinning of ClassSheets, a large variety of errors can be prevented that occur in many existing spreadsheet applications today.The presented ClassSheet approach links spreadsheet applications to the object-oriented modeling world and advocates an automatic model-driven development process for spreadsheet applications of high quality.

Process-Oriented, Flexible Composition of Web Services with UML

The composition of elementary web services to larger-scale services has become an important means to enhance e-business collaborations. If such composite web services can also integrate legacy components that are not yet provided as web services, the number of possible compositions is increased. Following a process-oriented approach, the compositions can be described as control- and data-flow between available web services and components. This paper discusses the Business Process Execution Language for Web Services (BPEL4WS), an existing service composition language, and proposes UML-WSC as an alternative, visual language. For the advanced description of service interfaces, UML-WSC extends the type system of the established Web Service Definition Language (WSDL).

Consistency-Preserving Model Evolution through Transformations

With model-based development being on the verge of becoming an industrial standard, the topic of research of statically checking the consistency of a model made up of several submodels has already received increasing attention. The evolution of models within software engineering requires support for incremental consistency analysis techniques of a new version of the model after evolution, thereby avoiding a complete reiteration of all consistency tests.In this paper, we discuss the problem of preserving consistency within model-based evolution focusing on UML-RT models. We introduce the concept of a model transformation rule that captures an evolution step. Composition of several evolution steps leads to a complex evolution of a model. For each evolution step, we study the effects on the consistency of the overall model and provide localized consistency checks for those parts of the model that have changed. For a complex evolution of a model, consistency can then be established by incrementally performing those localized consistency checks associated to the transformation rules applied within the evolution.