| Abstract: | 基於元件式的開發流程是多數開發團隊所採用的開發方式之一，而且也是目前最被接受的方法，此類型的開發必須基於前人所設計出來的元件之上而加上個人的需求設計，這些元件中存有前人的設計智慧與Domain Know-How於其中，以這些元件之上做個人需求設計可以快速的達到目前所要的需求，並且，無形中沿用了前人所設計系統架構，由此可知元件式的開發之所以會被廣泛接受並不無道理所在。 1997年由美國Carnegie Mellon University的軟體工程學院(Software Engineering Institute, SEI)所提出的CMMI開始，軟體工程品質的議題逐漸延燒，在品質的領域中，雖然很早之前就已有學者提出，但長久以來一直缺乏標準的規格被提出與公認，CMMI的出現為軟體品質的領域注入一道新血，開始有許多學者以CMMI為基礎研究軟體品質相關議題，舉凡軟體流程改善(Software Process Improvement, SPI)、Software Measurement/Metrics…等等都是其中的研究議題之ㄧ。 在本研究中所要討論的議題就是針對軟體品質中軟體度量/指標(Software Measurement/Metrics)為主，如同上述所講，元件式的開發雖然帶來了許多開發上的優點，舉凡避免重複開發元件而增快專案進度、Domain Know-How的採用、系統架構的延用…等，但是以品質的角度去分析，人們開始懷疑對於已存在元件的品質，會想要知道這些已存在元件的品質狀況如何，因此，在這方面的研究即是利用Software Measurement/Metrics的理論來達到這個目的；研究中就是從這方面的理論著手，針對目前應用相當廣泛的JavaBeans元件程式碼做度量，度量程式碼再使用(Reuse)的可行性，達到除了元件的再使用之外，也能夠做到程式碼的再使用，所以在本研究中基於軟體度量的理論並實作一軟體度量輔助工具，使用者可以透過此度量工具進行JavaBeans程式碼的度量，另外還可以藉由度量工具所提供的修改建議而進行適當的改善，以期望JavaBeans程式碼能夠符合再使用的設計方式。
Component-base software development process is the most accepted process. This kind of process is base on exist component and adds specify requirement into design. These components keep design technology and domain know-how inside. We can develop quickly base on this kind of component. In the other side, insidiously we extend original architecture. Therefore, because of these advantages causes component-base software development process widely used. When CMMI was proposed in 1997 by Carnegie Mellon University Software Engineering Institute in American, software quality issue has become more enthusiastic. Although software quality has already been addressed before, but it always lacks of standard specification for a long time. Till CMMI was proposed, many researchers start to research in this issue, such as “Software Process Improvement”, “Software Measurement”, “Software Metrics”, etc. In this paper, the subject we discuss is focus on Software Measurement/Metrics in software quality. As describe above, although component-base software development process has so many advantages such as reuse component, increase project schedule, adopt domain know-how, extend architecture, etc. But from the viewpoint of quality, people start to doubt about the existed component quality. They want to know the quality status of the component. Thus, we can use Software Measurement/Metrics theory to achieve this. We use this as our core theory in this paper. Because JavaBeans component are used widely, we aim at JavaBeans source code for our research target. Measuring the reusability of JavaBeans source code is good or not. We hope beside component reuse, we can even reuse source code. Further, we can modify source code by suggestion that provided from measurement CASE tool in order to map the reusability design of JavaBeans. |
