Study On The CMMI Software Process Measurement

The production of quality software on time and within budget has been a major concern of software industry for several decades. Applying systematic approaches in software development to control the development process can help produce product effectively and efficiently. And it is clear that organizations can best pursue systematic process improvement by integrating measurement and process.The software process measurement is the basic of checkup the deviation between actual performance and receivability performance, and it can get the foundation of improving the chance too at the same time. The organization of the software carries on the software process measurement, can gather to the tolerance data of each process performance, and analyze the performance of each process.In the thesis, based on CMMI framework, combining GQIM paradigm, an improved software process measurement modeling is put forward which is more available and customizable. First, we establish organization software process. According to goals of every sub-process in process model, we refine measurement goals. Aims to every measurement goal, questions are put forward. Based on questions, we define seven categories of indicators. At the same time we define three levels metrics construct which are indicator-derived metrics-basis metrics and establish indicators modeling based on the metrics construct. We detailed describe interpretation and operation about indicators, functions, metrics methods and counting rules etc. which assures metrics definers formalization. In other hand, indicators paradigm integrate data collection and analysis process, ensuring metrics result answer questions, consequently realization measurement goals.In order to measure and analyze the software production process, we need to collect enough data. But data collecting isn't our destination, we just use these data to identify and analyze the pattern, tendency and connection of process. In this article, we further analyze effect of software measurement on process improvement, apply Statistical Process Control technique to measurement analysis, establish software process capacity baseline.Finally, we use metrics data which was collected by above measurement paradigm and process capacity analysis arithmetic to establish an actual data model and test it to validate practicability and feasibility about above methods. At the same time, we analyze characteristic of the model on use and data model definition.