本研究嘗試針對特定應用領域的軟體系統，推導出一個能衡量出系統中程式與程式之間的相對複雜關係的複雜度衡量模式，並且嘗試驗證此模式的衡量結果在實務上能具有參考價值。 因此，在特定領域軟體系統的選擇上，本研究選定國內某銀行建置在IBM Mainframe平台上，以IBM Mainframe COBOL語言撰寫的信用卡應用系統之程式做為衡量標的，從中挑選數支樣本程式並利用本研究推導的複雜度衡量模式加以衡量，以求取出這些樣本程式之相對複雜度排序。接著，本研究亦請負責維護這些樣本程式的程式人員，針對同樣的程式依其本身對程式複雜程度的認知，排出相對順序。最後，本研究驗證這二個複雜度排列順序之間，有高度的正相關存在。 藉由與特定系統之程式開發人員本身，對於此系統中程式之複雜度認知的結果做比較，本研究證明了此衡量模式可應用在實務上之系統複雜度衡量上，並希望此衡量模式所產生的相對複雜度衡量結果，能對於軟體系統的一些特質的衡量，如開發人日工時的量測、系統錯誤率的量測等，有更進一步地幫助。 In this study, we try to define a measure method which is sensitive to many software characteristics for the complexity of special application software. Thus, we choose a credit card system of a bank in Taiwan. The system environment of this credit card system is IBM z/OS390. The computer language used to write this system is IBM Mainframe COBOL. We randomly select several programs from the system to be the sample. We use our method to measure these programs’ complexity first, and then obtain these programs’ relative complexity by sorting their complexity measures. To verify if our method is acceptable, we also request some programmers who are responsible for the development of this credit card system to sort those sample programs by complexity. Then we use the correlation coefficient test to measure the association between the two sorting results. The correlation test shows that our measure method is acceptable. We hope the measure method can be applicable in some management work, for example, man power allocation for software development, the prediction of fault rate of a software system, etc..
