Please use this identifier to cite or link to this item: http://dx.doi.org/10.18419/opus-10496
Authors: Koss, Patrick
Title: Evaluating and improving a scenario-based analysis method for service-based systems
Issue Date: 2019
metadata.ubs.publikation.typ: Abschlussarbeit (Bachelor)
metadata.ubs.publikation.seiten: 80
URI: http://elib.uni-stuttgart.de/handle/11682/10513
http://nbn-resolving.de/urn:nbn:de:bsz:93-opus-ds-105137
http://dx.doi.org/10.18419/opus-10496
Abstract: Modifiability and maintenance are important topics in software development. Many metrics, such as McCabe's cyclomatic complexity, therefore attempt to quantify the maintainability of software quality attributes, in particular the sub characteristic modifiability. Quantifying specific changes in systems is not an easy task. With scenario-based methods, certain changes can be described. However, this requires a high manual effort. A previously developed tool-supported method attempted to estimate the difficulty and complexity of changes by focusing on service-based systems and using a scenario-based assessment. Because the quantification of specific changes to a system cannot be fully automated, the tool-assisted approach has tried to be lightweight, so it does not require much manual effort. This work focuses on improving and evaluating the existing tool-supported method. We collected information about the technical background, then analyzed the existing toolsupported method and made a list of improvements. Some of them have been implemented. In a next step, we analyzed the now improved tool-supported method by conducting hands-on-interviews with seven participants to give a qualitative assessment, and by conducting a survey with 40 participants to provide a quantitative assessment. Our most important finding in the evaluation is that the survey participants found the lines of code estimation is neither very accurate nor applicable and therefore not suitable for our tool-supported method according to the participants of the survey. Story Points seem to be best suited as an effort estimation method for our five predefined estimation methods according to the participants of the survey. We found no correlation between the personal background of a participant, whether a participant has worked with scenario-based evaluation in real projects, or his familiarity with service-based systems and our effort estimation methods and our evaluation functions. Our findings from the hands-on interviews have been to improve the ease of use for creating scenarios, to better visualize the evaluation features and to determine which evaluation features were helpful and which not. After collecting a derived improvement list and implementing some suggestions, we believe that our tool-supported method can be used to estimate the effort of changes in a real service-based system, as participants, as our review notes, find the evaluation view and evaluation features helpful. Keywords: scenario-based evaluation method, lightweight, service-based systems, modifiability, maintainability, tool-supported method
Appears in Collections:05 Fakultät Informatik, Elektrotechnik und Informationstechnik



Items in OPUS are protected by copyright, with all rights reserved, unless otherwise indicated.