Quality assurance
Software quality, and measure factors
Quality Quality = + program quality software engineering software
Program quality and specific application scenarios most likely be the accuracy, quality query speed, concurrency, and security of the international.
Simple quality software engineering, it is a function of cost, time, visibility, including software development (software development can demonstrate as soon as possible), risk control, management factors, the delivery of quality, cost control, completion, internal quality (such as test the number of cases of use, the degree of automation of the test, daily build speed, efficiency of the automatic deployment of the system, code coverage, quality of a document, etc.)
Different standards to measure the quality of software engineering, in which a more mature theory is CMMI (Capacity Maturity Model Integrated, Capability Maturity Model Integration), practice has shown that the use of the model management project, not only reduces the cost of the project, but the project raised quality and on time completion rate.
CMMI is divided into five levels: initial level and management level, a clear level, quantitative management level, optimization level. Each level is a cornerstone of the higher level. It can be used to measure the ability to manage a project, can also be used to describe the maturity of an enterprise.
To achieve a certain degree of software quality is to pay costs, probably including prevention, assessment, internal fault, external fault, process analysis and improvement, improve their professional skills, technology and investment. Developers generally a little time each day to write code new features, additional time is spent on the cost of software quality.
Quality Assurance QA work
QA and testing are different, QA is to make the software meet the quality standards defined in advance and carried out all activities, including testing.
Everyone can participate in the QA work, but in the end there must be a role is responsible for QA it.
There are several software development process prone to doubt the following questions:
1, with a special test does not mean developers are not responsible for the quality, because the testers are not the most people understand the program, the developer is. When the self-test developer, an independent attitude to the test, rather than thinking about how to hide it
2, after the division of labor, each for their own role and performance optimization, local optimization rather than globally optimal happens
3, the test is not entirely responsible for testing, but also to understand the procedure
4. To clear division of labor, each character has his responsibility
Some within the company no independent testing role, this is because:
1, each character relatively strong
2, company-wide staff often use their products, demand can be analyzed by user feedback and real-time status, or have a special team to look for external bug
This team should train and arrange for a role:
