First, the experience on software quality assurance
1, the definition of good and bad software quality: The quality of the software primarily refers to whether it can meet the needs of users. Therefore, software quality assurance focused on "customer" is the core. Code phase: to write code to achieve customer needs, develop software stages: to reduce the risk and cost of the greatest degree, on the basis of the test phase to meet customer needs on: Bug does not meet the test software if there is customer demand.
2, the general concept of software quality assurance: quality assurance software includes many aspects, maintaining the quality of the quality of the preliminary procedure code, software engineering, software testing, software and later, small and complicated many aspects of software quality constitutes a good or bad the decisive factor. Therefore, software quality testing alone is not independent from the accuracy of the program code for comprehensive coverage of the beginning of the index to the software development process for risk control, cost control, internal software quality metrics completion to the final software testing, software quality is not so much that they are guaranteed by step, each step as it parallel and mutually inclusive.
3, concrete steps Software Quality Assurance division:
First, the software program = + Software Engineering
Available software program quality + quality = Quality Software Engineering
(1) guarantee the quality of the program: To be measured by the composite indicator multi-dimensional characteristics of the different needs of different software so comprehensive index is not the same.
(2) Quality Engineering Software Assurance: CMMI implementation
Second, if you are a QA project, what do you think your job responsibilities
-
1. The company responsible for carrying out the project's software quality assurance work.
-
2. The process of real-time understanding of the project, more with developers to communicate and exchange.
-
3. Develop project quality assurance program, the implementation of QA activities as planned, periodic reports project situation.
-
4. carry out the project development process to promote the reliability of activities to ensure delivery of products is stable and reliable.
-
5. The process of modifying products to monitor and supervision to ensure modify process control, quality and reliable.
-
6. From the above extraction procedure improved activity required to promote the continual improvement process.
-
7. Participation internal audit found that the point system improvements, and monitor project quality testers together.
-
8. involved in software testing, writing documentation.
Third, if you are a project manager, a full-time QA do you think your project needs? Or you can just have a Test? Who is responsible if problems arise how you define?
1, first of all talk about QA and Test differences and connections.
Test: the use of certain processes and tools, the software enables verification functions and features pre-designed, especially to verify that the code of conduct consistent with the provisions functional specification (spec) work processes and results are usually quantifiable.
QA: use a variety of means to ensure software quality software can help the team to achieve goals in all phases of software engineering, software teams are all activities to allow the software to meet the quality standards defined in advance and carried out, including testing.
Under this definition, Test can actually be said to be a proper subset of QA.
2, for QA:
I think full-time QA is very necessary
QA is the quality assurance of the entire software engineering, QA requirements throughout the entire product had (or project) the whole process, the whole process of supervision, the current progress of the project inspection, but also to complete the verification organization, organization process definition, risk management and other related processes. Each process is complete, require finishing phase of the work, this management quantitative, success and after the failure, analyze the reasons, at the same time accumulation of knowledge to improve the quality of subsequent products. So QA quality assurance on a piece of software is essential.
3, for Test:
I think it is not necessary to set up a full-time Test
(1) is contained as QA Test and QA if there is, it is not necessary to set up a full-time Test
(2) talk about existence is a misconception Test: Test all regulatory test work was carried out at the end of the whole project. In fact, if this is the case is not enough, when late in the project found the problem, the root cause is often some of the early decisions and project design, at this time, and then to modify more difficult. In fact, this requires testers from the project will begin to actively intervene to prevent the problem from the source. As a result, I feel that i work Test is not independent, whether it is in the early stages of software engineering (coding program) or in the latter part of the software design process, Test monitoring are essential. So if you want to avoid vulnerability STAGE out the entire pre-attached to correct things when late repair, will end at the beginning of the process from the Test has been associated to the process. If so, it is not as good as at the outset of this work by the developer. So that's the first reason, because Test should be monitored throughout the project.
（3）第三个原因就是参考资料主要谈论的问题，“不懂开发的人必然做不好测试”。因为没有人比开发人员更了解整个软件工程，在发现漏洞的时候也没有人比开发人员更快的找出问题所在，并找到解决的办法.
（4）第四个原因就是，书上质量保障这一章提到的：出现问题之后，开发人员和Test人员之间责任的推诿，和分工明确会造成画地为牢导致的局部最优和全局未必最优的情况。
4、出现问题之后责任的界定
首先作为项目经理，没有及时发现问题，很大一部分责任在我。我会反思自己在整个工程期间的工作，是因为自己没发现潜在问题还是出现了本不该出现的问题，若是前者，我会承担自己作为项目经理未做好本职工作的责任，若是后者，我会具体情况具体对待，细化究竟是哪个进程出现的问题即由哪个进程的负责人和我一起担责任。我觉得项目出现问题不会一个人的主要责任，同负责一个项目，应该一起面对所有的问题。