Art of Software Testing (summary)

The following section is a summary of the art of software testing

   The book is divided into the following sections: 1 economic and psychological tests and testing principles; 2 code review; 3 test base part; 4 development debugging and testing ideas; 5 Test the latest software applications;.....

  1, the economic and psychological testing and test principles

  Although software testing is a technical work, it is undeniable, he is also closely linked with human psychology. If you are sure in cognitive: Software testing is to find errors and execution of the program process. Well, first you in the design of test cases, there will be conscious to be able to design some of the problems found in test cases; then, in the implementation of test plans, will also find the wrong direction and goals as their own tests; of course, you will also found that software problems, improve software quality reflect their own value as a test program.

  Software testing will never find all the issues, in order to use minimal resources and cost find as many problems, we need to establish some strategies to deal with economic challenges. Including the black box, white box, all kinds of high-level test methods.

  Because the majority of software testing is a matter of psychology, so the authors conclude ten guiding principles of software testing, software development and software for knowledge testing. These principles are at work should be kept in mind, whether it is to do development or do the test, it can not only guide their better test based on these principles, but also can give inspiration in the development.

  Details can refer to: "Art of Software Testing (study notes 1)" ( https://www.cnblogs.com/chengabc/p/11215552.html )

  2, code review

  Code review include: code checking, code review, check the desktop and peer review. Work is still the most used in the code checking, code checks equivalent to the general code review. By contrast the method with the book and found that the company checks the code is not very formal. Irregular phenomenon as follows.

  (1) Code for checking

    Check the definition of the code is to form a group to read and inspection procedures. There are several problems,

    1. Membership is not strong sense of participation

    Most of the recommendations to improve the review process is indicated by the leadership, of course, led a lot of experience. This leads to a sense of involvement of other members is not strong, does not reflect the value of other members, if all the recommendations were put forward leadership, I can not be unconscious that the work of other members did not do.

    2. Lack of error list

    A review process does not check for errors in the list, with all the experience of members. If there is a check list of errors will increase the efficiency of the code inspection, but also allows each member to grow from the comparison list.

    3. Modify larger version

    Program issued version more, and not enough staff in the case, can only be reviewed by a leader or person in charge of eleven, that is, the desktop check. A good suggestion is submitted after the software version of the test, testers can test control before a list of errors, modify the point of the program review.

  (2) For the code review

    Walkthrough defined as: a team to conduct code reviews, difference is that members of the group of test cases to run in the program were deduced. The problem is that is not currently used such a review.

  Details can refer to: "Art of Software Testing (study notes 5)" ( https://www.cnblogs.com/chengabc/p/11331771.html )

  3, the test base part

    Test base comprising: a test case design method, the test method and the high-level unit testing methods.

  (1) test case design

    Test design method comprising: a black and white box box method.

    TESTING method commonly used are: equivalence partitioning method, boundary value analysis and error estimation process.

    White-box testing methods: statement coverage, decision coverage, covering condition determination / condition coverage, multiple condition coverage.

    The method is very simple, but if the program properly designed test cases using these methods in their work, is a very challenging task.

  Details can refer to: "Software Testing arts (reading notes 2)" ( https://www.cnblogs.com/chengabc/p/11254877.html )

  (2) Test Method Unit

    A single module test

    1. Analysis of the logical structure of module on the basis of the source;

    2. The white box test design method;

    3. Next viewing module specification description;

    4. TESTING method supplementary test.

    Testing multiple modules

    1. Incremental test

    2. Select the top-down or bottom-up test method

    3. Use automated tools to test for multiple modules

  Details can refer to: "Software Testing arts (reading notes 3)" ( https://www.cnblogs.com/chengabc/p/11260916.html )

  (3) High-level test method

    Higher level of testing is to reduce various types of software errors from end to demand information conversion between the code segments is inconsistent, resulting. Test higher level comprising: a module testing, functional testing, system testing, acceptance testing, usability testing. System testing which are more important considerations.

  Details can refer to: "Software Testing arts (reading notes 4)" ( https://www.cnblogs.com/chengabc/p/11297967.html )

  4, the development of debugging and testing ideas

  (1) Debugging thought

    1. Violence debugging method: see memory information, print the log, add a breakpoint Debug.

    2. induction debugging: through correct and incorrect data acquisition, data links between research, find out the cause of the error, and then solve the problem.

    3. deduction commissioning: According to collected data, a list of possible root cause, the cause can not be excluded based on the relationship between the data, and then solve the problem.

    4. Debug backtracking: errors in position, the reverse execution of the program to find the error location of program logic.

    Debugging process consists of two steps: positioning errors and correct errors. After the error changes, we should analyze mistakes, try to get some inspiration to improve coding and design.

  Details can refer to: "Art of Software Testing (reading notes 6)" ( https://www.cnblogs.com/chengabc/p/11335893.html )

  (2) Agile development and testing ideas

  In response to the growing demand for software, some developers propose ideas of agile development, agile development of the main features: customer participation, iterative development, test drive. Corresponding agile testing also needs to enhance the efficiency of software testing, agile testing features: customer engagement, automated testing, project follow-up.

  In a small project, you can speed up the development process by using extreme programming methods and limits of testing methods, and to ensure product quality.

  Details can refer to: "Art of Software Testing (study notes 7)" ( https://www.cnblogs.com/chengabc/p/11351334.html )

  5, test the latest software applications

  Chapter 10 of this book to explain the web test, Chapter 11 of this book explains the test mobile device, currently in the work are exposed to, it can be used to look after.   

    

Guess you like

Origin www.cnblogs.com/chengabc/p/11427909.html
Recommended