Search

COMPSAC 2008, Turku Finland, 28 July - 1 Aug. 2008, During a software system development, existing similar systems are useful because such systems and their documents help developers to understand and reuse problems and solutions of the new system. Especially, such helps are effective if developers are not familiar with the new system. In this paper, we present a supporting tool called “So/M” for an analyst to define software requirements. By using So/M, the analyst can easily refer functions of existing similar systems, their commonality and the relationships among them. So/M also enables the analyst to choose the candidate of requirements, and to generate the skeleton of requirements. During a comparative experiment, we have confirmed that So/M significantly helped analysts unfamiliar with a system to be developed., Article, Proceedings of the 32th Annual International Computer Software and App lications Conference. 255-262 (2008)

In a system development, the knowledge of a target business is very important factor for the success of a development. The needed part of such knowledge is different for each stakeholder. Therefore, we need a method to measure the level of understanding of a stakeholder and a method to categorize business knowledge. This paper proposes the metric of difficulty level in order to measure the business knowledge. Moreover, we performed the experiment to confirm the suitability of the proposed metric.

It is difficult to identify whether quality requirements are defined adequately or not, but there are few methods to support this kind of requirements analysis. In this paper, we propose a method based on software quality spectrum, that shows a ratio of quality characteristics embedded in a software engineering artifact, such as a requirements specification, a manual and so on. We assume similar kinds of software systems have similar spectrum, thus we can identify the adequacy of quality requirements for a new system by using spectrum of already existing similar systems. We confirmed the assumption above by analyzing actual software systems, i.e., web browsers and drawing tools.