Authors:
Michel dos Santos Soares
and
Jos Vrancken
Affiliation:
Faculty of Technology, Policy and Management, Delft University of Technology, Netherlands
Keyword(s):
UML, SysML, Petri nets, Empirical Software Engineering, Software Process Improvement, Action Research.
Related
Ontology
Subjects/Areas/Topics:
Enterprise Information Systems
;
Information Systems Analysis and Specification
;
Methodologies, Processes and Platforms
;
Model-Driven Software Development
;
Modeling Formalisms, Languages and Notations
;
Requirements Analysis And Management
;
Software Engineering
;
Systems Engineering
Abstract:
This article is about research performed by the authors into improving the Software Engineering process at an company that develops software-intensive systems. The company develops road traffic management systems using the object-oriented paradigm, and UML as the visual modeling language. Our hypothesis is that UML has some difficulties/drawbacks in certain system development phases and activities. Many of these problems were reported in the literature normally after applying UML to one project and/or studying the language’s formal specifications and comparing with other languages. Unfortunately, few publications are based on surveys and interviews with practitioners, i.e., the developers and project managers that are using UML in real projects and are frequently facing these problems. As a matter of fact, some relevant questions were not fully addressed in past research, mainly related to UML problems in practice. The purpose of this text is to report the main findings and the propo
sed improvements based on other methods/languages, or even considering UML diagrams that are not often used. The research methodology involved surveys, interviews and action research with a system developed in order to implement the recommendations and evaluate the proposed improvements. The recommendations were considered feasible, as they are not proposing to radically change the current situation, which would involve higher costs and risks.
(More)