Programming management principles and practices

Programming Management

Testing

Code

Basically, systems and applications are being developed to facilitate the traditional manual process. It is essential, therefore, that they must conform in the designed functionalities and specifications to effectively serve their purpose. Code testing is a critical element to this procedure to ensure that systems function well and that they meet user-requirements.

Manual

Manual code testing involves evaluation and assessment of program codes through manual process, whether the functionalities match those listed in the detailed design. In this test process, a series of codes are designed, developed, and embedded within the actual software or system codes to test the program's vulnerability. Normally, iterations are done to extensively test and validate code units. Following are some components of a manual code test process that can be expected in this section of a procedural manual.

A test-code design input design execution design

Manual code test process can also be implemented using debuggers. Some debuggers provide a step-by-step tracking of command lines and allow display of values that variables contain. In this way, should there be errors in the output of the program, the developer can trace where the error started to exist.

2. Automated

Automated test process is done with the use of test tools that automatically performs evaluation and assessment of a program code. This method implements validation of input, output, and program execution to test precision.

This section of a procedural manual may contain a description of the automated test tools that will be used. Similarly, the procedures to be followed in using the test tools must be also described in this section.

Unit Testing

Unit testing is the process of checking the units (i.e. modules, subprograms) of a software product. It involves individual checks and tests of every software component.

To carry out unit tests, different test methodologies may be used which can be performed either through manual or automated testing. Vector Software has the following list of test methodologies used in unit testing. This section of a procedural manual may use them as patterns to describe similar test cases.

Specification-oriented Testing

This method aims to check and ensure that all requirements are met in the software product, and that every software unit is functional and correct.

Implementation-oriented Testing

Test data are used in this method to check the various computational processes of the software. With test data, this method aims to discover any error that may occur during implementation of the software.

Error-oriented testing

All possible errors a software can possibly encounter are tested in this method. Its main focus is on software evaluation against potential errors.

Integration Testing

Integration testing is the process of checking all units of a software, when put together, must correctly function as one whole system. This type of test ensures that there is a proper communication between software units, and that there is a smooth flow of data between each other.

Integration test may involve different strategies. Following are examples of such integration test strategies.

Tests based on each particular process of the software

The various functionalities specified in the detailed design and specification upon integration of all components of the software is tested by this strategy. For instance, an Accounting system can be tested on its following functionalities: Record Search, Salary Computation, Record Maintenance

Tests based on test objectives (SQS Group)

This strategy involves testing of objectives common to software units. For instance, integration tests on functionalities that use one particular interface.

This section of a procedural manual may contain similar test cases of the above strategies.

Regression Testing

Errors found during testing may require modification of a software unit. Thereby, it is necessary that modified parts be re-tested to ensure that errors have been fixed and that the modified unit does not cause any error to other parts of the software. This process is called regression testing. Regression testing is also done when new codes are added.

In regression testing, the same set of test cases previously used to the software unit may be reuse. This method can provide the difference between the output of the unit before and after it was modified. Two approaches can be used in reusing test cases. The first approach is using all the established test cases previously applied to the software. The second is just selecting several test cases for rerun. This section of a procedural manual must contain the suitable approach to be used in regression testing, as well as the procedures in undertaking the test.