Abstract
Recently, small-scale project in a software development like as Web applications on client/server system are increasing with rapid progress of information infrastructure that is characterize by diffusion of a low-priced and high-speed Internet. These companies are pushed into a severe state by difficult operation of the small-scale project that continues increasing. Under such condition, these projects must achieve high-quality product to get user requirement with low-cost and short delivery time. Moreover, in order to realize high customer satisfaction, the company has spent large part of time of early stage of project to get user requirement. According to the textbook, early stage of a project is an important aspect of deciding success or failure of a project. Although enough time for risk response planning cannot be assigned, early stage of a project is hampered by the huge task of a requirement definition. In this paper, analyzing the 60 typical risks in software project that defined by Jones, following two ideas are proposed: (1)tracking the propagation pass like as neural-network for failure of a project is clearly shown, (2)according to analyze the propagation pass of project risk to project failure, the risk relation structure for easy risk response planning is provided. As the result that is mentioned above, accurate and easy risk identification method is proposed.
