軟體程序裁適技術之發展與應用

Development and Application of Software Process Tailoring Techniques

曾婉惠

指導教授 ： 范金鳳

元智大學/資訊學院/資訊工程學系/碩士(2003年)

摘要

現行軟體標準包含了繁多的執行程序及工作項目，但是每個專案都有不同的發展條件，因此並非標準中所有的工作項目都是必需執行。如何在有限的資源下決定哪些標準中要求的項目是符合成本效益已成為一重要的課題。
本研究的目的在於發展一套具有系統性(Systematic)的軟體工程標準裁適方法(Tailoring)。本裁適方法結合定性分析與定量分析，讓管理者判斷符合成本效益的執行項目。首先，我們利用UML的圖形表達將標準中的要求與專案執行中的實際面因素，接下來我們將UML轉換成BBN因果關係圖，利用BBN執行定性分析。再將BBN裡的因素，轉換成模擬變數，利用動態模擬執行定量分析。我們所提的定性與定量方法可用在事前專案計畫的裁適，亦可用在發展中的專案以進行監控，支援風險管理的決策，可降低專案發展的風險度。我們方法的最大特色在於根據標準進行具有系統性(systematic)的裁適方法，並應用此方法於特定的議題的裁適上，以證明其有效性。

關鍵字

裁適 ； 軟體程序 ； 貝氏信心網路 ； 監控

並列摘要

Software standard addresses many processes and activities in a software life cycle process. But every project has its own characteristics and its resources restriction, thus not all of the processes and activities discussed in software standards need to be executed. It has become an important issue to determine effective activities to carry out under limited resources of a software project.
In this research, we provided a systematic approach to tailoring the activities and documents required by on software engineering standards. In our method, we combined qualitative and quantitative analyses. First, we used UML to model industrial standards’ requirements along with pragmatic factors during the project development. Then, we converted the UML to BBNs for qualitative analysis of tailoring. We could further take BBN’s factors as simulation inputs to perform quantitative analyses of tailoring. Our qualitative and quantitative approach can be used for software activities and document tailoring before hand; it can also used to support the project risk management during project progress to reduce project risk. The major virtue of our method lies in its systematism based on software engineering standards. We have applied this method to specific tailoring issues to demonstrate its effectiveness.

並列關鍵字

Tailoring ； software process ； BBN ； monitor

參考文獻

[1] B.W. Boehm, ”Software risk management: principles and practices,”

[2] Marc Bouissou, Fabrice Martin, and Alain Ourghanlian, ”Assessment of a Safety-Critical System Including Software: A Bayesian Belief Network for Evidence,” IEEE Proceedings Annual Reliability and Maintainability Symposium, pp. 142-150,1999.

[4] James D. Arthur, Markus K. Groner, Kelly J. Hayhurst, and C. Michael Holloway, ”Evaluating the Effectiveness of Independent Verification and Validation,” IEEE, pp. 79-83 October 1999.

[5] Norman E. Fenton and Martin Neil, ”Software metrics, failures and new directions,” The Journal of Systems and Software, pp.149-157, 1999.

[6] K.G. Van Der Poel and S.R. Schach, ”A Software Metrics for Cost Estimation and Efficiency Measurement in Data Processing System Development,” Journal of Systems and Software 3, , pp. 187-191, Sep. 1983.