A user task design notation for improved software design.

Defect detection performance Player task notation Software design Software quality UML-Activity diagram

Journal

PeerJ. Computer science
ISSN: 2376-5992
Titre abrégé: PeerJ Comput Sci
Pays: United States
ID NLM: 101660598

Informations de publication

Date de publication:
2021
Historique:
received: 17 11 2020
accepted: 01 04 2021
entrez: 4 6 2021
pubmed: 5 6 2021
medline: 5 6 2021
Statut: epublish

Résumé

System design is recognized as one of the most critical components of a software system that bridges system requirements and coding. System design also has a significant impact on testing and maintenance activities, and on further improvements during the lifespan of the software system. Software design should reflect all necessary components of the requirements in a clear and understandable manner by all stakeholders of the software system. To distinguish system elements, separation of concerns in software design is suggested. In this respect, identification of the user tasks, i.e., the tasks that need to be performed by the user, is not currently reflected explicitly in system design documents. Our main assumption in this study is that software quality can be improved significantly by clearly identifying the user tasks from those that need to be performed by the computer system itself. Additionally, what we propose has the potential to better reflect the user requirements and main objectives of the system on the software design and thereby to improve software quality. The main aim of this study is to introduce a novel notation for software developers in the frame of UML Activity Diagram (UML-AD) that enables designers to identify the user tasks and define them separately from the system tasks. For this purpose, an extension of UML-AD, named UML-ADE (UML-Activity Diagram Extended) was proposed. Afterwards, it was implemented in a serious game case for which the specification of user tasks is extremely important. Finally, its effectiveness was analyzed and compared to UML-AD experimentally with 72 participants. The defect detection performance of the participants on both diagrams with two real-life serious game scenarios was evaluated. Results show a higher level of understandability for those using UML-ADE, which in turn may indicate a better design and higher software quality. The results encourage researchers to develop specific design representations dedicated to task design to improve system quality and to conduct further evaluations of the impact of these design on each of the above mentioned potential benefits for the software systems.

Identifiants

pubmed: 34084923
doi: 10.7717/peerj-cs.503
pii: cs-503
pmc: PMC8156997
doi:

Types de publication

Journal Article

Langues

eng

Pagination

e503

Informations de copyright

©2021 Ozcan et al.

Déclaration de conflit d'intérêts

The authors declare there are no competing interests. Eda Ozcan is employed by Vakifbank.

Références

Surg Innov. 2019 Dec;26(6):725-737
pubmed: 31370743

Auteurs

Eda Ozcan (E)

Vakifbank, Ankara, Turkey.

Damla Topalli (D)

Computer Engineering Department, Atilim University, Ankara, Turkey.

Gul Tokdemir (G)

Computer Engineering Department, Cankaya University, Ankara, Turkey.

Nergiz Ercil Cagiltay (NE)

Software Engineering Department, Atilim University, Ankara, Turkey.

Classifications MeSH