|
John Erickson, Keng Siau
|
|
Friday, 22 June 2007 |
UML’s complexity is regularly criticized by practitioners and researchers alike, who argue that such complexity is a considerable detriment to the adoption and use of UML in the field. Attempts have been made to assess and/or measure UML’s complexity in a number of ways. Erickson and Siau
proposed that a subset (kernel) of UML, composed of the most important constructs, could be equated with the complexity that practitioners face when using the modeling language. This research extends Erickson and Siau’s work by proposing a UML kernel in three application areas, real-time, webbased and enterprise systems. Compared to other modeling methods and languages, UML is very complex. As such, identifying a UML kernel will help in the training and usage of the language. In this research, we conduct a Delphi study using UML experts, to identify three UML kernels, and a nonspecific kernel, which are then combined into a single kernel.
UML’s complexity is regularly criticized by practitioners and researchers alike, who argue that such complexity is a considerable detriment to the adoption and use of UML in the field. Attempts have been made to assess and/or measure UML’s complexity in a number of ways. Erickson and Siau
proposed that a subset (kernel) of UML, composed of the most important constructs, could be equated with the complexity that practitioners face when using the modeling language. This research extends Erickson and Siau’s work by proposing a UML kernel in three application areas, real-time, webbased and enterprise systems. Compared to other modeling methods and languages, UML is very complex. As such, identifying a UML kernel will help in the training and usage of the language. In this research, we conduct a Delphi study using UML experts, to identify three UML kernels, and a nonspecific kernel, which are then combined into a single kernel.
[PDF] [BibteX]
Only registered users can write comments.
Please login or register. |
