Evaluating the Effectiveness of a Goal-Oriented Requirements Engineering Method

Abstract

As an attempt to answer the need for methods and tools in requirements engineering (RE) IntroductionThere has been strong evidence that RE needs proper engineering methods and tools, which are domain specific and comprehensive, in supporting major REOs, and that require very detailed tools associated with them, to produce high quality requirements, to save time and the effort of rework on requirements, and to reduce resources, such as the size of RE teams [1]. Moreover, there has been growing interest in GORE approach that is based on the identification of system goals and the transformation of these goals into requirements; it addresses concerns of why a certain goal is required, how it can be achieved and who is responsible for it in the system and/or the environment [2] [3] [22]. To address these issues, the research we are undertaking proposes to evaluate KAOS, a GORE method, and Objectiver, as associated tool for KAOS, for the following reasons: (i) the role of goals in RE is fundamental while they are a main part of use-cases in object-oriented and goal-oriented approaches; moreover, the goal notion is increasingly being used on current RE methods and techniques because there is a perceived inadequacy of the traditional systems analysis approaches when they are applied to complex software systems [2] [3] [5]; (ii) the KAOS method is the only method of the GORE family that pays special attention to the use of formal proof for model analysis; (iii) the Objectiver tool fully supports the KAOS methodology and is a commercially supported tool that can be relied on to help in the evaluation, and has been applied to a number of industrial problems and case studies already at Cediti-Belgium; (iv) the RE community is eager not only to understand KAOS and Objectiver, but also to be aware of how to improve them to be more supportive for the RE area and to extend the range of their success in solving most common RE complications; and (v) this research is believed to be the first research that related method-detected errors to REOs and objectives of the method itself, in order to give some processes or techniques as a guide to detect the diversion away from the REOs or their own objectives and get the development back on track. (ii) enabling a systematic derivation of requirements documents from requirements models; and (iii) improving the validation process, the quality of requirements documents and stakeholder communication [7]. The evaluation result will help to answer the research questions: (i) How well do KAOS and Objectiver meet the criteria established in the discipline of RE; and (ii) How well do KAOS and Objectiver achieve their own self-defined objectives. This paper describes our research, but, due to the page limitations, it presents only the part of the research that takes into account the major REOs using the first target problem and some of the primary results of the evaluation.

SCRAM–CK: applying a collaborative requirements engineering process for designing a web based e-science toolkit

“…The goals of a RE method can be used as the frame to evaluate the method [3]. This type of evaluation is qualitative, not statistical nor experimental.…”

Role Engineering with SKAOS for Systems Employing RBAC

“…It first determine the ultimate system goal, then keep decomposing to build the system object model, operation model, responsibility model, and delegation model etc, from the above series of models, the system requirement can be delivered [9]. However the KAOS method is too complicated to fulfill the role engineering needs, it takes to much effort.…”

Evaluating the use of Pareto Efficiency to Optimize Non-Functional Requirements Satisfaction in i* Modeling

“…Desde entonces, se han hecho esfuerzos en la comunidad de modelado conceptual para abordar este problema, por ejemplo, se han realizado evaluaciones empíricas para proporcionar información sobre el uso de modelos conceptuales por parte de los stakeholder's en la empresa [20] para la representación de las relaciones part-whole en el modelado conceptual [21], o para comparar diferentes lenguajes de modelado conceptual de datos [22]. En lo que respecta a la ingeniería de requisitos orientada a objetivos (goal-oriented requirements engineering), existen algunos trabajos que resaltan la importancia de la evaluación de este tipo de enfoques, Al Subaie y Maibaum [23] presentan una evaluación cualitativa del método KAOS y su respectiva herramienta Objectiver, este tipo de evaluación es subjetiva debido a que la información cualitativa de KAOS y Objectiver se basa en las observaciones de los evaluadores, usaron KAOS y Objectiver para abordar un problema objetivo para obtener informes y observaciones con el fin de evaluar la pertinencia, la exactitud, la trazabilidad y la comprensibilidad del enfoque. En cambio, en Teruel et al [24] estudiaron la aplicabilidad de tres técnicas utilizadas en la ingeniería de requisitos (casos de uso, orientados a objetivos y viewpoints) para especificar sistemas colaborativos, realizaron un análisis empírico para comparar las técnicas por medio de una variación del método DESMET [25], utilizado para evaluar métodos y herramientas en ingeniería de software.…”