Logo des Repositoriums

Auflistung nach Schlagwort "maintenance"

1 - 7 von 7
  • Konferenzbeitrag
    Architecture Recovery from Fortran Code with Kieker
    (Softwaretechnik-Trends Band 43, Heft 1, 2023) Jung, Reiner; Schnoor, Henning; Gundlach, Sven; Hasselbring, Wilhelm
    Scientific models are software systems, which are key to understand and assess a range of challenges, such as climate change mitigation. However, such models are usually developed over decades. To support program comprehension for software maintenance and restructuring, we designed an architecture recovery process for Fortran-based scientific models utilizing Kieker 4 C to collect call traces at runtime. Furthermore, we derive structural information from the recovered architecture. In this paper, we present our analysis process and some results from analyzing three scientific models. Additionally, we describe how to use the information obtained by our analysis to identify possible optimizations of the scientific models.
  • Konferenzbeitrag
    Automated maintenance of a software portfolio
    (Softwaretechnik-Trends Band 25, Heft 2, 2005) Veerman, Niels
    We report on an automated mass maintenance project. We updated an entire software portfolio automatically. The updated portfolio was accepted by the company and taken into production.
  • Konferenzbeitrag
    Generic Performance Measurement in CI: The GeoMap Case Study
    (Softwaretechnik-Trends Band 43, Heft 1, 2023) Reichelt, David Georg; Krauß, Hannes; Kühne, Stefan; Hasselbring, Wilhelm
    Continuously developed industry projects with medium budget often focus on functional correctness, but not on optimal performance. This is partially caused by the lack of easily available approaches and tooling that check for performance changes. The tool Peass-CI examines performance changes by measurement and analysis of the duration of unit tests. We present a case study of establishing a continuous performance engineering process in GeoMap, a Spring based tool that allows to analyze the real estate market for real estate service and market experts. In the continuous performance engineering process, we monitored performance changes that happened during six months and derived performance improvements by code reviews and load test execution. We found that (1) continuous performance measurement gives detailed insights into performance changes and (2) performance improvements by source code changes are reproducible using performance measurement of unit tests in Peass-CI.
  • Konferenzbeitrag
    Impact of augmented reality guidance for car repairs on novice users of AR: a field experiment on familiar and unfamiliar tasks
    (Mensch und Computer 2020 - Tagungsband, 2020) Hoffmann, Clemens; Büttner, Sebastian; Prilla, Michael; Wundram, Kai
    The use of augmented reality (AR) guidance is seen as an opportunity to address the growing complexity of industrial tasks. Previous research showed benefits of AR for different industrial tasks especially for novice users, while other research suggests that AR was not superior to other means for novices. However, there is not much work that looks at the relation between initial exposure of users to AR (that is, if users have never used AR before) and different types of tasks. In this paper, addressing the field of car maintenance and repair, we look into the question of how AR support impacts the performance in familiar and unfamiliar task if the AR user has never used AR before. By running an experiment under field conditions, we investigate whether the familiarity of a specific repair task has an impact on the performance under AR guidance compared to a traditional repair guideline. Our experiment reveals interesting insights. First, we show that familiarity and routine have an important impact on adherence to (all) repair guidelines, which should be regarded in future studies. Second, despite its novelty and the corresponding added time to deal with AR, we found that guidance via AR worked better for unfamiliar tasks. This shows the potential of AR for guidance of industrial tasks in practice, and it brings up design suggestions for the implementation of this guidance in practice.
  • Konferenzbeitrag
    Program Comprehension and Developers’ Memory
    (INFORMATIK 2019: 50 Jahre Gesellschaft für Informatik – Informatik für Gesellschaft, 2019) Krüger, Jacob; Wiemann, Jens; Fenske, Wolfram; Saake, Gunter; Leich, Thomas
    In this extended abstract, we summarize our paper “Do You Remember This Source Code?”, published at the International Conference on Software Engineering 2018 [Kr18]. We discuss implications of our results on forgetting in the context of program comprehension, providing a more contextual perspective on our results compared to the original paper and a previous abstract [Kr19].
  • Konferenzbeitrag
    Understanding How Programmers Forget
    (Software Engineering and Software Management 2019, 2019) Jacob Krüger; Jens Wiemann; Wolfram Fenske; Gunter Saake; Thomas Leich
    This extended abstract is based on our paper “Do You Remember This Source Code?”, published at the International Conference on Software Engineering 2018 [Kr18]. We summarize and discuss our results on programmers’ memory and forgetting. To this end, we focus on reverse engineering of software, which was the primary context in which we conducted this work.
  • Konferenzbeitrag
    Using Interaction Logs for Creation and Maintenance of Trace Links
    (Softwaretechnik-Trends Band 40, Heft 1, 2020) Hübner, Paul; Paech, Barbara
    Trace links between requirements and source code are beneficial for many software engineering tasks during development, e.g. maintenance, program comprehension and reengineering. It is necessary to keep trace links up-to-date for the continuous usage of the links during a project. Thus, maintenance of links along with changes in linked artefact is essential. Existing automatic trace link creation approaches are not capable to continuously provide and maintain links with high precision for direct usage by developers during a project. We therefore extended the commit linking by data captured from developers work. In our completely automatic approach ILCom, we use interaction logs recorded in an IDE while a developer is working on code files to implement a requirement. To assign the recorded interaction logs automatically to the requirement worked on, ILCom uses the issue IDs of requirements provided by developers in commit messages. This enables continuous link creation after each commit without manual effort from the developers. We evaluated our approach in several studies and showed that our ILCom approach is the first trace link creation and maintenance approach with very good precision and recall which does not require any manual work from the developers besides the common commitbased linking. In our presentation we introduce ILCom and the results of the evaluation studies. Further we sketch how maintenance of existing links along with changes in linked artefacts can be integrated in ILCom.