Software Test Summary (a) - Basic Concepts

Software Test Definitions

Software testing is manually operated automatically by the software under test system software or static checking tools, it is an object under test or validate the software system meets the requirements , or to clarify the difference between the actual output to the desired output.
Software testing is always around demand expanded.

Four step software development

1. Define requirements (Principal)
2. Analysis of demand (both sides)
3. Implementation requirements (developer)
4. Check the demand (testers)

Typically, the software development life cycle, the test should be involved as early as possible. Currently prevalent involved in the analysis of the requirements phase.

Four steps of software testing

1. Plan: Develop a test plan based on the document (requirements specification, preliminary design documents or detailed design documents) and project planning.
2. Design: Design test cases and test methods.
3. Implementation: According to test configuration testing environment, design-driven modules, and finally run the test, the test results, look for defects.
4. Assessment: Quality assessment test work to assess the quality of the tested software system.

The basic steps for dynamic testing manual

1. Providing the measured object
2. Prepare related expected (ie, various documents)
3. Design test cases
4. Set up a test environment
5. Run the test case
6. Test execution results
7. Record execution
8. The report found defects
9. Regression testing

The basic steps of manual static test

1. Providing the measured object
2. Prepare related expected
3. Reviewing code, documentation
4. The report found defects

We can see the static test compared to less dynamic test procedure that looks easier to implement, but the static test requirements for higher technical personnel, will be discussed later in detail.

Test Case

Since the actual software testing should be balanced between quality, cost and time of the three, the actual software testing is often based on the lowest risk, highest efficiency divide and rule principle to design. Thus, we need to have test cases arise.
Refers to test small test unit to represent the demand , which describes the expected output of the user, describing the actual results of the implementation of the system. Mainly includes three parts: input, output and testing environment.

Software Testing Model

Software Testing Model main waterfall model, V model, W and H model model four.

Waterfall model

Traditional waterfall model is only in the testing process as needs analysis, outline design, detailed design and coding of a later stage, this division usually leads to late discovery of defects resulting in high repair costs.

V model

V model improved the shortcomings of the traditional waterfall model, on one side of the development phase, starting with defined business needs to start, and then converts the demand for the software specification, and then switch to the outline design and detailed design, and finally encoded as program code. On one side of the test execution phase, the first unit testing and integration testing, system testing, and finally the acceptance tests that form different levels of software testing (level), and corresponds to a respective stage of the development process.

W model

W model consists of two "V" formed by overlapping. Wherein a "V" represents a development process, including needs analysis, general design, detailed design, coding implementation, module integration, system construction, system installation phases. Another "V" represents a test process, including testing requirements, designed to test summary, detailed design testing, unit testing, integration testing, system testing and acceptance testing and other activities. Software testing and test activities at all stages of the development process, respectively.
It is now widely used software testing model, pointing out that it should be involved in software testing requirements analysis phase.

H model

H model emphasizes the test is divided into test preparation and test execution two different stages, as long as due process led to progress in other test-ready point of the place, this time, can not be completed as long as the test preparation, test execution activities can or need to carry out, in the H model among emphasize testing is a completely separate model , it is possible to carry out cross and other processes, but also allows us to perform a test as soon as possible.

Software testing is not complete

Software testing is not complete, there are risks. Software testing can identify defects in the software, the software can not guarantee that no defects. A good test is to find software defects so far not found.

Test Method Evaluation Criteria

Test methods should be designed in the shortest time with the least human, found in favor of the largest and most serious flaw. So therefore the method has the following five evaluation criteria:

1. Accurate: stronger targeted test (number of test cases less)
2. Complete: To fully test coverage, bug-free (coverage requirements covering the risk)
3. No redundancy: needs and risks associated test cases should be differentiated
4. Simple: the test method should be simple and easy to implement
5. Ease of commissioning: the difficulty of locating small defects