Authors:
Leandro F. Silva
1
;
Edson OliveiraJr
1
and
Avelino F. Zorzo
2
Affiliations:
1
Informatics Department, State University of Maringá (UEM), Maringá, Brazil
;
2
Pontifical Catholic University of Rio Grande do Sul, Porto Alegre-RS, Brazil
Keyword(s):
Experiment, SMartyModeling, Software Product Line, Tool, UML.
Abstract:
Variability modeling in UML-based Software Product Lines (SPL) has been carried out using basically the UML Profiling mechanism for a diverse of theoretical approaches. However, there is no UML-based SPL life cycle supporting tool, which takes advantages of the UML standard diagrams in a controlled environment exclusively dedicated to it. Users usually adopt general-purpose UML tools to model variability. Its drawback is no control over data regarding SPL models, especially on variability. With such control, one might, for instance, use different visualization techniques to show SPL/variability information, inspecting/testing SPL models and data, apply metrics, and configure specific products. To provide an environment with these characteristics, we developed SMartyModeling. We evaluated its feasibility based on two studies: one qualitative supported by the Technology Acceptance Model (TAM), and one experiment comparing SMartyModeling with Astah. The first study aided to establish as
sumptions on how to improve the environment. We then, stated hypotheses to be tested in a comparative experiment. Thus, we identified aspects related to the automation of the SPL concepts, the number of errors and the difficulties in modeling SPLs. Hence, we measured effectiveness and efficiency of SMartyModeling over Astah. General results provide preliminary evidence that SMartyModeling is feasible for further developing.
(More)