Digital Library of Gesellschaft für Informatik e.V.
Not a member yet
40758 research outputs found
Sort by
Introducing BPMNGen: An LLM-based Conversational Framework for BPMN 2.0 Process Model Generation
No full textBusiness process modeling is a crucial task for organizations to document, analyze, and optimize their business processes. However, creating process models requires modeling expertise, aggravating this task for non-experts. Recent advancements in Natural Language Processing and artificial intelligence offer new possibilities for automating the generation of process models from natural language descriptions. This doctoral research aims to address these emerging possibilities by developing BPMNGen, an LLM-based conversational framework that enables users to generate and iteratively evolve BPMN 2.0 process models through interactive prompts using natural language input. The research follows the Design Science Research Methodology by developing and evaluating the conversational framework in a structured manner. The main goal of this research is to make process modeling with BPMN 2.0 more accessible and convenient for a broader audience
A formative evaluation: co-designing tools to prepare vulnerable young people for participating in technology development
No full textParticipatory processes are key for designing technology solutions but challenging since target groups rarely have technological expertise. They are, however, experts in their lives. Vulnerable young people are even more challenged when asked to participate. A toolkit approach can introduce non-tech-savvy individuals to digital technology. This paper presents the process of (co-)designing a toolkit for vulnerable young people and social workers. The evaluation provides insights into what components are needed to prepare for participatory technology development. The toolkit can be seen as a mobile laboratory that supports technology development and builds on considerations from other research and design processes. The toolkit was (co-)designed in two phases. An initial version based on other prototypes and the state of the art was evaluated according to feedback from social work professionals. The revised version was evaluated by both professionals and vulnerable young people, leading to refinement and final redesign. The final toolkit, focusing on the smart home domain, was evaluated through participant observation and qualitative content analysis. The results show how feedback, inductively derived in four categories, led to the design of the toolkit with three modules relevant to engaging vulnerable young people in technology design: experience, understand and co-design
Towards Green Software Metrics for Sustainable Software Development
No full textDespite the growing attention of Green Software Metrics (GSM) in facilitating sustainable software development within contemporary IT to environmental sustainability, the practical application and effectiveness of GSM tools remain insufficient. This research aims to address these shortcomings by identifying sustainability metrics, developing a robust methodology for their precise measurement, and evaluating existing tools, across a range of modern runtime environments, including bare metal servers, virtual machines, containers, and cloud-based systems. This topic is highly relevant to the performance community, as carbon emissions must be integrated alongside traditional metrics like response time, throughput, and resource utilization. The research aims to provide developers with actionable insights, guiding the adoption of sustainable practices and reducing the environmental impact of IT operations. Experiments will help to understand this distinction better, as well as provide a guideline for determining which tool might apply best in a given scenario. Furthermore, this will assist in determining the extent to which the aforementioned metrics facilitate the advancement of sustainable software development
On the Impact of Requirements Smells in Prompts: The Case of Automated Traceability
No full textLarge language models (LLMs) are increasingly used to generate software artifacts, such as source code, tests, and trace links. Requirements play a central role as they are often used as part of the prompts to synthesize the artifacts. However, the impact of requirements phrasing on LLM performance remains unclear. This paper investigates the role of requirements smells—indicators of potential issues like ambiguity and inconsistency—when used in prompts for LLMs. We conducted experiments using two LLMs focusing on automated trace link generation between requirements and code. Our results show mixed outcomes: while requirements smells had a small but significant effect when predicting whether a requirement was implemented in a piece of code, no significant effect was observed when tracing the requirements with the associated lines of code. These findings suggest that requirements smells can affect LLM performance in certain SE tasks but may not uniformly impact all tasks
Ein Vergleich von Attract-Funktionen in SEE
No full textIn diesem Beitrag stelle ich einen Vergleich von Methoden zur Vorschlagsauswahl für Zuordnungen während der Reflexionsanalyse vor, der anhand automatischer Experimente auf sechs Java Open-Source-Projekten in der Visualisierungsplattform SEE durchgeführt wurde. Der Vergleich, die Visualisierung sowie eine Benutzerstudie über den Einfluss der Vorschläge gingen aus meiner Masterarbeit hervor
TAV 50 - 50 TAV
No full textRückblick auf 50 Jahre bzw. 50 Treffen der Fachgruppe
»Test, Analyse und Verifikation von Software
Entwicklung eines Conversational Interface zur Steuerung von SEE und Abfrage von Graphdatenbanken
No full textDieser Beitrag fasst meine Bachelorarbeit zur Erweiterung der Anwendung SEE (Software Engineering Experience) zusammen. SEE ist ein System zur kollaborativen Softwarevisualisierung, bei dem Anwender als Avatare eine virtuelle Welt betreten, in der Software-Projekte als sogenannte Code-Cities dargestellt werden. Zu Beginn meiner Arbeit verfügte SEE bereits über einen Assistentin-Avatar. Die Interaktion mit diesem sollte um ein Conversational Interface erweitert werden. Des Weiteren soll die Assistentin „intelligenter“ werden. Durch eine Repräsentation der Software-Projekte als Graph in einer Graphdatenbank sollte die Möglichkeit geschaffen werden, mit Hilfe der Assistentin dort hinterlegte Software-Metriken abzufragen. Durch Sprachbefehle sollen weitere Funktionen in SEE gesteuert werden. Nach Fertigstellung eines Prototypen wurde eine Nutzerstudie durchgeführt. Die Studie hat gezeigt, dass die Verwendung des Conversational Interfaces je nach Aufgabe einen positiven Effekt auf die Bearbeitungszeit haben kann
Rezension zu „How to Win Client Business – When You Don’t Know Where to Start“ von Doug Fletcher
No full textALPINE: Abstract Language for Pipeline Integration and Execution
No full textWhen working with data, it is essential to ensure data quality and clean data of errors. This is usually done with a data cleaning pipeline. The execution of such a pipeline is possible with a variety of tools. However, to increase reusability, there should be a way to describe pipelines independently of technology. So far, there is no such technology-independent description. This paper therefore presents ALPINE, a language for describing data cleaning pipelines. This abstracts from the concrete implementation. After introducing the individual components, the usage is illustrated using a running example