Collecting Feedback during Software Engineering Experiments
Empirical Software Engineering
Flexibility in research designs in empirical software engineering
EASE'08 Proceedings of the 12th international conference on Evaluation and Assessment in Software Engineering
Empirical-WebGen: a web-based environment for the automatic generation of surveys and experiments
EASE'08 Proceedings of the 12th international conference on Evaluation and Assessment in Software Engineering
On the Reuse of Heterogeneous IPs into SysML Models for Integration Validation
Journal of Electronic Testing: Theory and Applications
Hi-index | 0.00 |
Several processes have been proposed for the transitionfrom functional requirements to an object-orienteddesign, but these processes have been subject to littleempirical validation. A use case driven developmentprocess is often recommended when applying UML.Nevertheless, it has been reported that this process leadsto problems, such as the developers missing somerequirements and mistaking requirements for design. Thispaper describes a controlled experiment, with 53 studentsas subjects, conducted to investigate two alternativeprocesses for applying a use case model in an object-orienteddesign process. One process was use casedriven, while the other was a responsibility-drivenprocess in which the use case model was applied as ameans of validating the resulting class diagram. Half ofthe subjects used the modelling tool Tau UML Suite fromTelelogic; the other half used pen and paper. The resultsshow that the validation process led to class diagramsimplementing more of the requirements. The use casedriven process did, however, result in class diagramswith a better structure. The results also show that thosewho used the modelling tool spent more time onconstructing class diagrams than did those who used penand paper. We experienced that it requires much moreeffort to organize an experiment with a professionalmodelling tool than with only pen and paper.