Assessing the Suitability of UML for Capturing and Communicating Systems Engineering Design Models

Author(s)
Organization(s)
Attachment(s)
FileSize (KB)MIME typeLanguage
SuitabilityOfUMLForSE_2002.pdf283.07application/pdfEnglishDOWNLOAD!
Abstract

Over the past thirty years, systems engineers have developed proven methodologies and design models to address the challenges of developing complex systems. With the rise of software-intensive systems, some organizations now support the adoption of the Universal Modeling Language (UML) as a standard language not only for software engineering, but also for systems engineering. This paper evaluates the suitability of UML as a systems engineering language in terms of (a) the capabilities which the existing language provides to practicing systems engineers, and (b) whether UML would improve the integrity of the design model. The findings are that UML has many logical inconsistencies in both the use cases and class models, and would therefore degrade the integrity of the system design model.

Introduction

Large projects throughout the defense, aerospace, and commercial domains require systems engineers to create, update, and document the design of large complex systems. To address these challenges, the systems engineering community has developed and refined standards, methodologies, and the basis for a common language centered around a series of fundamental tasks and practices. Though the community recognizes the need to interact with a diverse set of domain specialists and stakeholders in order to successfully design and implement complex systems, the techniques and methodologies developed have focused at the systems level and are not domain-specific.

As software replaces hardware as the dominant component in many complex systems, there is a renewed focus on the interface between systems and software engineers. A controversy has arisen within the systems engineering community, as some believe that the best approach for improving communication is to standardize on a common language for design models across the communities. The Universal Modeling Language (UML), well documented and accepted as a standard within the software engineering community, is being hailed by some organizations as the language that should be adopted for systems engineering. Other groups resist the adoption of UML within systems engineering and endorse the continued evolution and application of proven systems engineering methodologies, techniques, and design models

The purpose of this paper is to evaluate UML as a systems engineering language in terms of (a) the capabilities that the existing language provides to practicing systems engineers, and (b) whether UML would improve the integrity of the design model.

Rating
0
Referenced standards and/or methodologies
Visit also