"Software Test" class Note 03

There are many ways to test black boxes, mainly divided into the following:

1. Boundary Value Analysis method

2. Causality Diagram method

3. Decision Table method

Boundary Value Analysis Method:

Definition: A large number of software testing practices have shown that failures often occur on the boundary of a defined field or range, rather than within it. Designing test cases specifically for processing near the boundary of the detection is usually a good test result. Therefore, the boundary value analysis method is a very useful black box test case method, it has a strong ability to find fault.

Condition: The boundary is a special case. The program is correct when handling a large number of intermediate values, but errors can occur at the boundary. The boundary condition is the edge condition where the operating bounds of the software plan are located.

Boundary value and equivalence class relationship: Equivalence class division is divided by boundary. 　　The equivalence class test case selection should choose the boundary value as the test case. Method: (Here are only two examples of variables, the multivariable can be extended according to the two variables) 1.　　By making all variables normal (nom), only one variable is taken to the minimum (min), slightly above the minimum (min+), slightly below the maximum (max-), and Max (max).　　2. That is, by making all variables normal (nom), only one variable is taken to be slightly below the minimum (min-), Minimum (min), maximum (max), and slightly higher than the maximum (max+). 3. Robustness (robust) boundary value test. 4. Worst case Test (worst cases). Causality diagram method: definition: If there is no connection between program input, it is a more effective method to use equivalence class partition and boundary value analysis.　　If there is a relationship between the input, for example, the constraint relation, the combination relation, the equivalence class division and the boundary value analysis are difficult to describe, the test effect is difficult to guarantee, so it is necessary to consider a test method which is suitable for describing the combination of various conditions, producing multiple corresponding actions, and the causality diagram is presented in this context. 　　Represents the symbol: 1. 　　2.3. 　　4.5.  6.7. Decision Table Method: (strictest) Definition: A decision table is a table that lists all the combined values of all inputs as conditions and the corresponding output values. It is able to enumerate complex problems in all possible cases, concise and avoid omissions.　　Therefore, the decision table can be used to design a complete set of test cases.　　Structure: a decision table is usually composed of a conditional pile (Conditions), a condition item (Condition alternatives), an action pile (actions), and an action item (action Entries) 4 parts. Steps:

(1) List all the condition piles and action piles.

(2) Determine the number of action rules (condition items).

(3) Fill in the conditions.

(4) Fill in the action item and get the initial decision table.

(5) Simplifying decision tables and merging similar rules.

The above is the introduction of three methods of Heihe test. Most of the content comes from the teacher's courseware.

Self-Summary:

1. For the boundary value method, I think all the programs should be in the line boundary value test. Write for several years code, most of the errors appear in the value of the field, whether overflow, etc., that is, the boundary value problem. It is hoped that in future procedures the boundary should be considered as well as possible.

2. Decision-making is one of the most rigorous test methods in black-box testing and is an essential step, but the decision-making table method is somewhat difficult, so it is very important to pay attention to it.

"Software Test" class Note 03