- 學位論文
同仁審查之流程與成功關鍵因素探討─以台灣證券交易所系統設計審查為例
Study for critical success factors of peer reviews in software development process – a case at system design review policy of Taiwan Stock Exchange.
摘要
近年來我國政府為提升國內軟體開發組織的競爭能力，開始推展CMMI (Capability Maturity Model Integration) 認證制度。CMMI共分為5級，在第三級 (Level 3) 流程中，必須完成軟體「驗證」之「同仁審查(Peer Reviews)」特定目標。軟體的同仁審查方法國外已研究多年，但依最正式的方法執行，會耗費相當大的人力及時間成本，對國內佔大多數的中小型軟體開發組織幾乎無法執行。如採非正式的方法，則無法達到最高效率，因此本研究探討同仁審查活動關鍵因素，分析各因素之重要性順序，並以台灣證券交易所同仁審查流程為個案研討對象，案例研究結果可做為未來執行非正式同仁審查之參考。 本研究先以文獻探討蒐集同仁審查重要成功關鍵因素(Critical Success Factors)，再使用「德菲法(Delphi Method)」問卷，請相關專家學者確認並評估關鍵因素重要性。並以「分析層級程序法(Analytic Hierarchy Process, AHP)」分析專家回應資料，根據分析結果提出個案流程改善意見，最後再另請另組專家就前述建議事項，評估其可行性及績效。根據專家評估認為同仁審查組成要素之重要性，依序為程序、人員與技術，而依重要性序之前五項關鍵因素，分別是完善的檢查表、規劃的程序、開發流程中應規劃同仁審查流程、審查人員會議前充分的準備及設定同仁審查目標。最後，本研究所建議的流程及技術亦可為其他組織開發軟體時，執行同仁審查之重要參考依據。
並列摘要
In the past couple of years, the ROC government has adapted the CMMI (Capability Maturity Model Integration) system to promote domestic software development competitiveness. The CMMI is divided into five levels, and it requires a "peer review" verification of the developed software at its third level. The peer review system has been studied for a long time. If it is realized in a correct way, the formal peer review will takes a lot of time and human resource, and its costly procedure cannot be accepted by smaller software companies. However, if the peer review is realized in an informal way, the quality of software cannot be guaranteed. Therefore, this study analyzes the critical success factors in the peer review, and rearranges these factors according to their importance. We propose the steps of the review as a case, and bring up ways to improve according to our conclusion, which can be used as a reference for small to mid size businesses. In this research, we will first discuss documents, and collect peer review processes and important application methods to see what factors make peer review successful. The second phase is to approach experts familiar with the peer review, and give them questionnaires using the Delphi method, and collect information about the critical success factors. The third phase involves using the Analytic Hierarchy Process to analyze the views of experts and bring up suggestions for improving the process flow according to the conclusion. In the fourth and final phase, we interviewed the experts and discuss whether the new procedures are feasible.