III. HOW TO CONTROL PRODUCT QUALITY OF MEASURE ANALYSIS
A. Software Metrics
It is a very difficult issue of what to measure throughout the development process in the end. Measure the real users of the data is not only the project team, it is more of a manager. When management do not care about the organization's process capability, software metrics will lose direction and goals; project group and the management does not care about the organization's process performance metrics lost the role and significance of its existence. These two aspects, we must first determine the measure of the primary objective of the activity. Measure (discussed below "metrics") must be determined by the business goals of the enterprise for a specific problem.
B. Production Analysis
Many organizations start with the first area of software quality is the product software measurement. Most of the data can be obtained from the documentation of software problem reports and software programs, such as the number of software defects; the size of software; development costs; success by the number of test cases. Characteristics of the software product metrics mainly deal with product attributes, defects and other product-related data. Quality control person (QC) position code wrong location, design weaknesses and test through software product metrics. These metrics can
make QA personnel identify the product and make a analysis for product. After long-term analysis of the product quality control person can get product information and data used in the analysis process.
C. Process Analysis
In the software development process, no matter what kind of software metrics is including basic software metrics. The basic process of software metrics are as follows:
1) Measure of commitment:
According to the demand of the technology and management of the software development process on software metrics to determine the measure of the target, select the metrics to determine the focus of the implementation of software process metrics, which is targeted the first step to enhance the software metrics, but also senior managers to participate in decision-making and provide the appropriate resources an important part.
2) Measurement plan:
The development of software metrics program based on established software measurement goals, according to the software development technology, management, processes and problems. Officially confirmed in the plan related issues and attributes of products, processes, roles, responsibilities and resources, written for the implementation of software metrics, planning, feasibility and resources to support the guarantee.
3) Measures to implement the measure:
measure objects on software development projects, products and processes based on software metrics program implementation. By measuring the collection, storage, analysis of effective software metrics, metrics and analysis results are used to control and improve the software process.
4) Assessment of measure:
Evaluate the software measurement process by itself metric, metrics, processes, metrics and measure objects, measure the effectiveness of assessment found the measure of job problems, summed measure the operating assets and measure operations to improve the program.
5) Measures to improve in the follow-up measure of job:
According to the measure ofjob improvement plan is implemented, it will improve the import of the program the next time the software measurement process. It not only improved a simple repetition of operations in the horizontal direction, but also based on the experience lessons learned spiraling process and the effectiveness of software metrics in the software development process unfolded. Once based on business objectives and identified the need to measure what the content, methods and responsible person, you can begin data collection.[3]
D. Selecting the right software metrics
Determine measurable goals and select the appropriate metrics do measure. We determine the measurement goals, commonly used Goals-Questions-Metrics(GQM) method. The GQM method proposed by Dr.Victor Basili in the University of Maryland and his assistant, it told the organizations or institutions to which data should be collected. The GQM method implicit assumption is that every organization, the project has a series of goals to be achieved; In order to achieve each goal to answer a series of questions in order to know whether the target is achieved; each of them can be found by a complete and satisfactory answer quantified. GQM process is as follows:
1) a list of goals;
2) a set of problems to describe the target;
3) the definition of metrics need to answer these questions;
4) the development of data collection and analysis mechanisms;
5) collect, verify, analyze data and take the right action;
6) post-mortem analysis of the data in order to assess whether the goal is the same, and to provide recommendations for its improvement after;
7) for stakeholders to provide feedback.
Measurement goals is the development of information needs, possible sources involved such as estimating the parameters of the project plan, the implementation of the supervision of the state of the project, the experience of established management objectives, business plans, formal requirements or contractual obligations, other projects or organizational level entities and process improvement plans.
IV. CONCLUSION
Through the metrics of the paper, we increase the rate of code residual defect from 5% to 1%. At the same time, we increase the software project effort estimation accuracy from 60% to 80%. In addition to the above aspects, there are other factors to consider the software product quality management, such as risk management, change management and configuration management. In fact, the software developed by the U.S. Software Engineering Institute (SEI) Capability Maturity Model (CMMI) and ISO 9001 standard focus on quality and process management. Aspects of the organizational structure, mature foreign software companies generally have a separate QA (Quality Assure) department and develop department, it is independent and responsible for overseeing the implementation of the process. However, the implementation of any specific project should formulate
an appropriate quality management program, it cannot be born to move rigidly, and these need to project the accumulation of experience and learning new knowledge constantly.