In most UML-based methodologies, the analysis tasks include mainly modeling the functional requirements using use cases, and modeling the problem domain using a class diagram. Different methodologies prescribe different orders of carrying out these tasks, and there is no commonly agreed order for performing them. In order to find out whether the order of these analysis activities makes any difference, and which order leads to better results, we carried out a comparative experiment. Subjects were asked to create the two analysis models for a certain system in two opposite orders, and the qualities of the produced models were then compared. The results of the experiment reveal that the class diagram is of better quality when created as the first modeling task, while no significant effect of the analysis order was found on the quality of the use cases. We also found out that analysts prefer starting the analysis with data modeling.
|Number of pages||12|
|Journal||CEUR Workshop Proceedings|
|State||Published - 1 Dec 2006|
|Event||11th International Workshop on Exploring Modeling Methods for Systems Analysis and Design, EMMSAD 2006 - Held in Conjunction with the 18th Conference on Advanced Information Systems, CAiSE 2006 - Luxembourg, Luxembourg|
Duration: 5 Jun 2006 → 9 Jun 2006
ASJC Scopus subject areas
- Computer Science (all)