The Study of Establishing Quantitative Performance Indicator for Software Process Improvement

• Main Authors: Tsung-mu Wu, 吳宗穆
• Other Authors: Sun-jen Huang
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/r5artu

碩士 === 國立臺灣科技大學 === 資訊管理系 === 95 === Capability Maturity Model Integration (CMMI) has become a widely-adopted model for improving software development or maintenance processes all over the world. Although the performance results of CMMI-based software process improvement (SPI) revealed by Software Engineering Institute (SEI) website are positively significant, SEI does not provide any performance indicator in its CMMI website and thus organizations that reported their performance results defined their own performance indicators and the associated calculation equations. Therefore, these revealed performance results cannot be used for the benchmark comparisons. Nowadays, many domestic organizations are adopting CMMI for their process improvement, but there are still no quantitative performance indicators for domestic organizations to compare their SPI performances with other organizations. In order to provide consistent and objective performance indicators for assessing the benefits of CMMI-based process improvement, it is urgently important to establish the standard quantitative performance indicators and the associated metrics sets for the domestic CMMI-adopted community. This thesis has established the Quantitative Performance Indicators (QPI) for the CMMI-based process improvement. By referencing goal-driven software measurement (GDSM) methodology and six performance categories in CMMI website, the proposed QPI includes 29 software metrics which are grouped by cost, schedule, quality, customer satisfaction, return of invest, and productivity. The detailed operation definition including 30 data items for each of QPI is also provided in this thesis. This study has also performed questionnaire investigation from domestic scholars and experts and further held a group meeting to verify the completeness, feasibility and applicability of the proposed OPI. Finally, these 29 QPI are classified into three levels, namely required, expected and informative, in each of maturity levels (ML 2 to ML5) by taking account of the availability of base measures required in a performance indicator and the degree of its ease of use. This helps the use of the QPI for evaluating the benefits of the CMMI adoption more effectively.