TY - GEN
T1 - Stepwise refinement of software development problem analysis
AU - Kobayashi, Tsutomu
AU - Ishikawa, Fuyuki
AU - Honiden, Shinichi
PY - 2016/1/1
Y1 - 2016/1/1
N2 - The Problem Frames approach has attracted attention because it enables developers to carefully analyze problems in a reasonable manner. Despite that this approach decomposes a problem into sub problems before the analysis is conducted, developers are still faced with a complex analysis when they consider interactions between the various sub problems. Moreover, progressive evolution of requirements is important for flexible development. In this paper, we propose methods to analyze multiple abstraction layers of a problem. Our methods help developers to construct abstract versions of a problem and find relationships between abstract problems and concrete problems. Moreover, our methods support refinement of arguments such that the properties of the abstract problem are preserved in the concrete problem. Therefore, our methods enable developers to divide up arguments into multiple abstraction layers and thus mitigate the complexity of argumentation. We carried out preliminary experiments on abstracting problems and constructing reasonable arguments. Our methods are expected to enable developers to analyze problems in a reasonable manner with less complexity and thus make problem analysis easier.
AB - The Problem Frames approach has attracted attention because it enables developers to carefully analyze problems in a reasonable manner. Despite that this approach decomposes a problem into sub problems before the analysis is conducted, developers are still faced with a complex analysis when they consider interactions between the various sub problems. Moreover, progressive evolution of requirements is important for flexible development. In this paper, we propose methods to analyze multiple abstraction layers of a problem. Our methods help developers to construct abstract versions of a problem and find relationships between abstract problems and concrete problems. Moreover, our methods support refinement of arguments such that the properties of the abstract problem are preserved in the concrete problem. Therefore, our methods enable developers to divide up arguments into multiple abstraction layers and thus mitigate the complexity of argumentation. We carried out preliminary experiments on abstracting problems and constructing reasonable arguments. Our methods are expected to enable developers to analyze problems in a reasonable manner with less complexity and thus make problem analysis easier.
UR - http://www.scopus.com/inward/record.url?scp=84997545125&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84997545125&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-46397-1_37
DO - 10.1007/978-3-319-46397-1_37
M3 - Conference contribution
AN - SCOPUS:84997545125
SN - 9783319463964
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 488
EP - 495
BT - Conceptual Modeling - 35th International Conference, ER 2016, Proceedings
PB - Springer-Verlag
T2 - 35th International Conference on Conceptual Modelling, ER 2016 held in conjunction with Workshops on AHA, MoBiD, MORE-BI, MReBA, QMMQ, SCME and WM2SP, 2016
Y2 - 14 November 2016 through 17 November 2016
ER -