On Functional Size Measurement in the Early Software Development

浅谈早期软件开发中的功能规模衡量

基本信息

批准号：
17500022
负责人：
KUSUMOTO Shinji
金额：
$ 2.48万
依托单位：
Osaka University
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (C)
财政年份：
2005
资助国家：
日本
起止时间：
2005 至 2007
项目状态：
已结题

项目摘要

In this research, we aim to establish a method to estimate functional size in the upper phase of software development. At first, we examined an automatic functional size measurement (use case point) from use case model. In order to measure use case points, it is necessary to extract actors and use cases and then judge complexity of them. But, it is conducted manually and so some accidental error could be observed. To cope with the problem, we have developed an automatic use case measurement tool. We have also applied the tool to actual use case models and examined the difference between the value by the tool and one by the specialist. As the results, UCPs measured by the tool are similar to ones by the specialist. Next, we examined an automatic functional size measurement (function point) from source code. It has been reported that since function point measurement involves judgment on the part of the counter, differences for the same product may occur even in the same organization. Also, if an organization is going to adopt FP, it is necessary to measure it from the past software developed there. Here, under several conditions, we propose a method to extract data and transaction functions from Web application using static analysis based on LFPUG method and develop a function point measurement tool. Then, we have applied it to several Java programs and examined the difference between the FP values obtained by the tool and those of an FP measurement specialist. As the results, the number of data and transaction functions extracted by the tool is similar to ones by the specialist.

在这项研究中，我们旨在建立一种在软件开发上层阶段估计功能尺寸的方法。首先，我们检查了用例模型的自动功能尺寸测量（用例）。为了衡量用例点，有必要提取参与者和用例，然后判断其复杂性。但是，它是手动进行的，因此可以观察到一些意外错误。为了解决问题，我们开发了一个自动用例测量工具。我们还将工具应用于实际的用例模型，并检查了该工具与专家的值之间的差异。结果，该工具测量的UCP与专家相似。接下来，我们从源代码检查了一个自动功能尺寸测量（功能点）。据报道，由于功能点测量涉及计数器的判断，因此即使在同一组织中，同一产品的差异也可能发生。另外，如果组织要采用FP，则有必要从过去开发的过去软件中进行测量。在这里，在某种情况下，我们提出了一种基于LFPUG方法的静态分析，从Web应用程序中提取数据和事务功能的方法，并开发功能点测量工具。然后，我们将其应用于多个Java程序，并检查了该工具获得的FP值与FP测量专家的FP值之间的差异。结果，该工具提取的数据和交易功能的数量与专家相似。