Contributions published at Requirements Engineering (Martin Glinz)

Many organizations providing products with common features wish to take advantage of that similarity in order to reduce development and maintenance efforts. Their goal is to move from a single-system development paradigm towards a product line approach. However, this transition is not trivial and requires a systematic scoping phase to decide how the product line should be defined, i.e. what products and features should be included and thus developed for reuse. Currently available product line scoping approaches require huge upfront investments in the scoping phase, consuming a lot of time and resources. Our experience has shown that small and medium enterprises require a lightweight approach to decide how the existing products are related to each other so that their potential for reuse can be estimated more easily. In this paper we present a conceptual solution and early tool support enabling companies to semi-automatically identify similarity within existing product configurations.

The number of available cloud services has recently grown considerably. Therefore, consumers demand new methods for choosing the appropriate cloud services for their needs, whereas providers require dedicated ways to correctly elicit requirements from very heterogeneous consumers. In this poster, we present the StakeCloud community platform. This will act as a cloud resources marketplace, allowing consumers to input their needs and providing them with matching cloud services. Moreover, in case the needs are not met, they can be communicated as new requirements to cloud providers. Our solution will enable easier resource identification and requirements communication in cloud systems, thus supporting both consumers and providers.

Product Lines can be used to make it possible for the customer to arrange an individual product configuration. But the gain of new requirements for the product lines by the vision of the end user at the moment is only possible with high costs. In this paper we present a method to combine the product configuration and the extraction of new requirements by the end user in a single process. In this process the user describes his requirement in natural language. The program analyses the text and compares it with the existing models of product lines. The program asks then matching questions to get a valid product configuration. In the scope of this paper a first prototype of this solution has been implemented and evaluated. The evaluation resulted in a positive finding that recommends continuing the studies with this method.

In Sofware Engineering, it often happens that programers implement new software requirements directly in the source code and afterwards release a new version of the software. Thereby the update of the requirements specification is forgotten very often. Because this work is very difficult and enourmously time-consuming in hindsight, we create in this thesis a new approach as well as a corresponding tool to make this work easier. Several already existing source code-comparison algorithms are analyzed regarding their suitability for our purposes. Based on the gained insights a new approach is developed, implemented and tested. The developed tool extracts keywords out of two following releaseversions of a program. These keywords describe the changes that have been made in the source code. The extracted keywords shall help the developper to identify differences in the source code without big effort and give indication of which parts of the requirements specification need to be adapted.

In early requirements engineering phase pen and paper, flipcharts and whiteboards are still the favorite choice to sketch ideas and diagrams. Existing software solutions often lack the flexibility required for modeling diagrams, which would allow the freehanded drawing. The sketches often have to be reworked for the documentation, which in turn can be faulty and time consuming. This bachelor thesis shows with the help of an application for mobile tablet devices an alternative for the ad hoc modeling of requirements. It includes the development of a tool that combines simplified modeling of diagrams with the approach of the freehanded drawing. A key aspect of the implementation is the development of a sketch recognition algorithm that recognizes the drawings and thus interactively supports the modeling. Afterwards, the implementation is tested with the help of a usability evaluation.

Smart distribution networks shall improve the efficiency and reliability of power distribution by intelligently managing the available power and requested load. Such intelligent power networks pose challenges for information and communication technology (ICT). Their design requires a holistic assessment of traditional power system topology and ICT architecture. Existing analysis approaches focus on analyzing the power networks components separately. For example, communication simulation provides failure data for communication links, while power analysis makes predictions about the stability of the traditional power grid. However, these insights are not combined to provide a basis for design decisions for future smart distribution networks. In this paper, we describe a common model-driven analysis framework for smart distribution networks based on the Common Information Model (CIM). This framework provides scalable analysis of large smart distribution networks by supporting analyses on different levels of abstraction. Furthermore, we apply our framework to holistic survivability analysis. We map the CIM on a survivability model to enable assessing design options with respect to the achieved survivability improvement. We demonstrate our approach by applying the mapping transformation in a case study based on a real distribution circuit. We conclude by evaluating the survivability impact of three investment options.

Determining the trade-off between performance and costs of a distributed software system is important as it enables fulfilling performance requirements in a cost-efficient way. The large amount of design alternatives for such systems often leads software architects to select a suboptimal solution, which may either waste resources or cannot cope with future workloads. Recently, several approaches have appeared to assist software architects with this design task. In this paper, we present a case study applying one of these approaches, i.e. PerOpteryx, to explore the design space of an existing industrial distributed software system from ABB. To facilitate the design exploration, we created a highly detailed performance and cost model, which was instrumental in determining a cost-efficient architecture solution using an evolutionary algorithm. The case study demonstrates the capabilities of various modern performance modeling tools and a design space exploration tool in an industrial setting,provides lessons learned, and helps other software architects in solving similar problems.

[Context and motivation] Traditionally, requirements are documented using natural language text. However, there exist several approaches that promote the use of rich media requirements descriptions. Apart from text-based descriptions these multimodal requirements can be enriched by images, audio, or even video. [Question/Problem] The transcription and automated analysis of multimodal information is an important open question, which has not been sufficiently addressed by the Requirement Engineering (RE) community so far. Therefore, in this research preview paper we sketch how we plan to tackle research challenges related to the field of multimodal requirements analysis. We are in particular focusing on the automation of the analysis process. [Principal idea/results] In our recent research we have started to gather and manually analyze multimodal requirements. Furthermore, we have worked on concepts which initially allow the analysis of multimodal information. The purpose of the planned research is to combine and extend our recent work and to come up with an approach supporting the automatic analysis of multimodal requirements. [Contribution] In this paper we give a preview on the planned work. We present our research goal, discuss research challenges and depict an early conceptual solution.

Context and motivation Quality requirements are a main driver for architectural decisions of software systems. Although the need for iterative handling of requirements and architecture has been identified, current architecture design processes do not provide systematic, quantitative feedback for the prioritization and cost/benefit considerations for quality requirements. Question/problem Thus, in practice stakeholders still often state and prioritize quality requirements before knowing the software architecture, i.e. without knowledge about the quality dependencies, conflicts, incurred costs, and technical feasibility. However, as quality properties usually are cross-cutting architecture concerns, estimating the effects of design decisions is difficult. Thus, stakeholders cannot reliably know the appropriate required level of quality. Principal ideas/results In this research proposal, we suggest an approach to generate feedback from quantitative architecture evaluation to requirements engineering, in particular to requirements prioritization. We propose to use automated design space exploration techniques to generate information about available trade-offs. Final quality requirement prioritization is deferred until first feedback from architecture evaluation is available. Contribution In this paper, we present the process model of our approach enabling feedback to requirement prioritization and describe application scenarios and an example.

Today, the purpose of a model is often kept implicit. The lack of explicit statements about a model's purpose hinders both its creation and its (re)use. In this paper, we adapt two goal modeling techniques, the Goal-Question-Metric paradigm and KAOS, an intentional modeling language, so that the purpose of a model can be explicitly stated and operationalized. Using some examples, we present how these approaches can document a model's purpose so that this model can be validated, improved and used correctly.

Although end-users are key stakeholders in many software projects, current RE tools and approaches fail to properly address them. This thesis proposes the use of end-user communication tools - namely social network sites - in order to provide end-users with a collaborative space for requirements elicitation and negotiation. Based on the well-established platform Facebook we show what RE activities can be supported with the existing social network elements and present a Facebook-supported RE approach. The approach was then used in the context of 3 studies where Facebook users actively participated in RE activities, moderated by RE students. The studies show that end-users actually were actually to use the tools in order to brainstorm and discuss requirements, as well as to prioritize them - without further training. The quality and quantity of the requirements gathered suggest that crowdsourced RE can and should be enabled via social network sites.

Systematic reuse of artifacts and a clear understanding of the variability within a product family are key success concepts within diverse industrial domains. Nevertheless, there are still many open issues regarding adapting and tailoring of software product line engineering approaches to specialized domains. The nature of ERP systems would suggest the application of product line techniques, but the limitations and constraints within this domain makes this a challenging task from the viewpoint of a partner company. In this paper we discuss ERP domain constraints and provide first conceptual solutions on how to adapt and extend software product line techniques for this particular context. Furthermore, we present a first tool prototype to support sales consultants at ERP partner companies.

null

Recent approaches to end-user requirements elicitation focus on rich media descriptions, allowing smartphone users to express their needs using images, contextual information, and their everyday language. As RE experts cannot be expected to handle corresponding submissions entirely manually in large-scale scenarios, ways of automating the preliminary analysis of multimodal end-user requirements are needed. The present work explores the possibility of grouping similar requirements by means of a dedicated clustering pipeline. In a detailed analysis of existing submissions, 22 challenges posed to automated clustering have been identified and classified. Based on these findings, this report outlines a conceptual solution for enabling the clustering of multimodal end-user requirements. Both the classification of identified challenges and the conceptual solution are intended to inform further research towards large-scale end-user requirements elicitation with mobile tools.

Variability modeling is a major issue in requirements engineering for software product lines. Existing state-of-the-art approaches heavily leverage the principle of separation of concerns by specifying the requirements model in multiple separate diagrams (for example, the UML) and using orthogonal variability modeling to describe commonality and variability (for example, feature models). Mapping-based approaches are further used to accurately specify the variability’s impact on the requirements model. However, this makes variability-related requirements engineering activities unnecessarily cumbersome, since the specification of variable features is not handled as a primary concern, but rather scattered over many separate diagrams. This thesis presents a new approach that builds on different premises, which neither require variability mappings nor an orthogonal variability model. Instead of separating the requirements model into multiple diagrams, a fully integrated requirement modeling language and tool support for view generation are used. Instead of using an orthogonal variability model and a specific variability mapping approach, a compositional approach is used. On this basis a new, fully-fledged product line requirements modeling approach has been developed. The approach is parsimonious in the sense that it aims at extending an existing language as little as possible. It also allows a fine-grained specification of cross-cutting features and their functional dependencies, when needed. It allows the requirements and variability model to be visualized in a single view and more abstract views at arbitrary levels of abstraction to be generated. It provides novel support for product derivation that can visualize both a decision’s impact on the product’s functionality and on other variability decisions in the same diagram and tool. It continuously verifies the satisfiability of the model and allows advanced automated analyses such as constraint propagation, based on Boolean satisfiability (SAT) solving. In addition, it provides advanced support for variability model creation and evolution by allowing a straightforward, semi-automated extraction and composition of any identified variable feature. The empirical validation presented in this thesis is four-fold. First, a constructive tool implementation proves technical feasibility. Second, the modeling of several real-world examples along with state-of-the-art solutions shows practical feasibility. Third, a rigorous performance analysis verifies that SAT solving scales well for models of this new type, which proves that the presented automated variability analysis solution is feasible. And fourth, a recent real-world case study compares the practical performance of the presented approach with an industry- strength and state-of-the-art solution and shows considerable benefits of the presented approach. This thesis contributes a complete description of this new approach, which is mainly based on the A DORA requirements and architecture modeling language. The presented approach is of a general nature, however, and we expect our empirical results to yield equally encouraging data also with any other language that satisfies the approach’s prerequisites. We hope that this integrated approach to requirements modeling (or conceptual modeling in general) and variability modeling will soon also be applied with other modeling languages. The presented work will possibly lead to an emergence of new types of tools that can also visualize existing product line models (for example, specified with UML and feature modeling) in a new, integrated and flexible manner, as presented in this thesis. This could profoundly change and improve the way engineers and analysts visualize and deal with variability in software models in the future. vii Variabilitätsmodellierung ist eine wichtige Angelegenheit in der Anforderungsanalyse von Software Produktlinien. Existierende Ansätze mit heutigem Stand der Technik stützen sich stark auf eine Separierung der Anforderungsmodellierung in verschiedene Diagramme (typischerweise unter Verwendung von UML) und benützen orthogonale Variabilitätsmodellierung zur Beschreibung von Gemeinsamkeiten und Variabilität (zum Beispiel Feature Modelle). Zuordnungsbasierte Ansätze werden weiters verwendet um die Variabilität in der Anforderungsmo- dellierung präzise zu spezifizieren. Dies macht Aktivitäten im Zusammenhang mit Variabilität in der Anforderungsanalyse unnötig umständlich, da die Spezifikation der Variabilität nicht als grundlegendes Element realisiert ist, sondern vielmehr verteilt über verschiedene Diagramme beschrieben wird. Diese Dissertation präsentiert einen neuartigen Ansatz, der weder Zuordnungen noch ein zusätzliches orthogonales Variabilitätsmodell benötigt. Anstatt das Anforderungsmodell auf mehrere Diagramme zu verteilen, wird eine vollständig integrierte visuelle Anforderungsmodellie- rungssprache zusammen mit Werkzeugunterstützung zur Sichtengenerierung verwendet. Und anstatt ein orthogonales Variablitätsmodell zusammen mit einem Zuordnungsbasierten Ansatz zu verwenden, wird ein kompositionaler Ansatz eingesetzt. Auf dieser Basis wurde ein neuartiger und vollständig ausgeprägter Ansatz zur Anforderungsmodellierung von Software Produktlinien entwickelt. Der Ansatz ist sparsam, da er darauf zielt mit einer kleinstmöglichen Erweiterung einer bestehenden Sprache auszukommen. Dennoch erlaubt er eine sehr genaue Spezifikation von querschneidenden variablen Features und deren funktionalen Abhängigkeiten, wenn dies gebraucht wird. Er erlaubt eine integrierte Darstellung des Anforderungs- und Variabilitätsmodells und die Generierung von abstrakten Sichten auf beliebigen Abstraktionsstufen. Weiters erlaubt er eine neuartige Herangehensweise zur Produktableitung, welche die Auswirkungen einer Entscheidung sowohl auf die Funktionalität des Produkts als auch auf die weiteren Entscheidungen im gleichen Diagramm und Werkzeug visualisieren kann. Die logische Erfüllbarkeit des Modells wird laufend geprüft, was auch eine breite Palette an automatisierten Analyseoperationen erlaubt, die auf Boolescher Erfüllbarkeitsauswertung (SAT) aufbauen und zum Beispiel die automatische Propagierung der Auswirkung von bestimmten Entscheidungen auf die restlichen Entscheidungen erlaubt. Der Ansatz erlaubt weiters auch eine fortgeschrittene Art der Erstellung und Weiterentwicklung von Variabilitätsmodellen durch einfache, semi-automatische Extraktion und Komposition ausgewählter variabler Features. Die empirische Validierung dieser Dissertation gliedert sich in vier Teile. Zuerst beweist eine konstruktive Werkzeugimplementierung die technische Machbarkeit des Ansatzes. Als Zweites zeigt die Modellierung von realen Beispielen im Vergleich mit existierenden Lösungen am Stand der Technik die praktische Tauglichkeit des Ansatzes. Als Drittes zeigt ein rigoroser Leistungstest, dass die präsentierte automatisierte Analyse des Variabilitätsmodells gut skaliert und ebenso praxistauglich ist. Und als Viertes zeigt eine aktuelle Fallstudie, dass der praktische Einsatz des präsentierten Ansatzes im Vergleich zu führenden Werkzeugen mit aktuellem Stand der Technik auch bedeutende Vorteile birgt. Diese Arbeit präsentiert eine vollständige Beschreibung dieses neuen Ansatzes, welcher hauptsächlich basierend auf der Anforderungs- und Architekturmodellierungssprache A DORA beschrieben wird. Der Ansatz ist jedoch von genereller Natur und wir erwarten ähnlich vorteilhafte Ergebnisse auch für andere Sprachen, die alle nötigen Voraussetzungen für diesen Ansatz erfüllen. Wir hoffen, dass dieser integrierte Ansatz zur Anforderungsmodellierung (bzw. zur konzeptuellen Modellierung im Allgemeinen) und Variabilitätsmodellierung auch bald in anderen Modellierungssprachen zum Einsatz kommt. Die präsentierte Arbeit wird möglicherweise zu neuartigen Typen von Werkzeugen führen, welche auch existierende Modelle von Software Produktlinien (welche zum Beispiel mit UML und Feature Modellierung spezifiziert sind) auf eine neue, integrierte und flexible Art und Weise visualisieren können, wie in dieser Dissertation präsentiert wird. Dies könnte die Art und Weise wie Ingenieure und Analysten in Zukunft Variabilität in Software Modellen visualisieren und behandeln grundlegend verändern und verbessern.

Due to significant industrial demands toward software systems with increasing complexity and challenging quality requirements, software architecture design has become an important development activity and the research domain is rapidly evolving. In the last decades, software architecture optimization methods, which aim to automate the search for an optimal architecture design with respect to a (set of) quality attribute(s), have proliferated. However, the reported results are fragmented over different research communities, multiple system domains, and multiple quality attributes. To integrate the existing research results, we have performed a systematic literature review and analyzed the results of 188 research papers from the different research communities. Based on this survey, a taxonomy has been created which is used to classify the existing research.

The private banking and thus the financial advisory process is changing, the quality of the current advisory services is low and the diffusion of tablets like the iPad is progressed throughout the year. This provides the background and motivation to investigate the improvement of the financial advisory process with the usage of an iPad. A new conception was designed and implemented with low- and high-fidelity prototypes. Experimental evaluations show that we could maintain the improved transparency and high acceptance and satisfaction of the tabletop system. Furthermore, we increased the advisor's perceived ability to control the encounter. In return, the client's desire for more influence increased as well.

Software Traceability is a field of research that becomes more and more important in the context of modern agile software development models. Since manual traceability is an expensive and time consuming task, several traceability tools have been developed in recent years to support the user through automatization. In this thesis four traceability tools have been applied on a traceability benchmark. Based on twelve artifact scenarios, their performance has been measured and compared by the two measurements recall and precision. Furthermore the runtime of each tool has been measured as the perceived time that a tool needs to generate the links. Depending on the artifact scenario the tools show higher or lower performance, whereas Retro has demonstrated the best performance in many scenarios.

People have many creative ideas, but only a few of these ideas are realized and lead to innovation. Good ideas often fail because they are not shared between innovators and stakeholders, hence are unlikely to be realized. Consequently, many opportunities are missed to excite customers and to gain a competitive advantage. This paper proposes an innovation process that uses online social networks to lower the hurdle to sharing ideas. The process leverages diffusion effects of social networks while supporting the generation, evaluation, consolidation, and implementation of innovative ideas with lightweight activities. The process is illustrated and discussed using an application example. Although we are focusing on innovation regarding software-intensive systems, we foresee that the discussed process has the potential to be applied to other domains as well.