Computer Aided Software Testing (CAST) tools can be used at all stages of the test process.

Not all of the tools actually run the program. Some of them are used to analyse the code before the code is even compiled. Further it must be remembered that a tool may have more than one use or capability.

Documentation Control and Management
Computer Aided Software Engineering (CASE)
Program Code
Management
Runtime Execution

Documentation Control and Management
These tools enable automated verification and validation of requirements models such as consistency checking and animation. Thus finding ambiguities and inconsistencies.

Computer Aided Software Engineering (CASE)
These tools are used to interface with an external CASE repository or to manipulate live or test data held in an existing database or file system.

Comparison tools
Detecting differences between actual and expected results are the aims of these tools. Using "comparators" for comparing GUI objects or bitmaps. Usually the tester can mask off areas of the screen that are not to be compared.

Test data preparation tools
Test data is prepared in existing databases. This data is then used to drive automated testing. Typical uses of this technology is for keyword or action word testing.

Test Design Tools
generate test cases from a specification held in a CASE tool repository or from requirements formally specified within the tool. Alternatively test cases are designed following an analysis of the code. Definition available on Test Design Techniques.

Program Code
Monitors the quality of source code and and executable code from a program design specification.

Static Analysis
Examining code, provide the information with these tools. They seek to give objective measures of characteristics such as McCabes Cyclomatic Complexity Metric and path coverage. See more on Static Analysis

Dynamic Analysis
These tools provide runtime information on the state of executing software. Monitoring allocation, use of, and de-allocation of memory leaks, unassigned pointers are tasks undertaken by these tools.

Coverage Measurement
Coverage tools provide objective measures of structural test coverage when tests are executed. Statement or branch coverage statistics are provided in a log file after execution.

Test Harnesses and Drivers
Where software to be tested involves an interface to another component that has not yet been built, a test harness can be used. Alternatively to run a series of controlled automated scripts.See more Dynamic Analysis

Debugging Tools
Developers are able to reproduce bugs and investigate the state of software. This can be done line by line. Execution can be halted at any time to check variables. A trace log can be used to record the code execution path.

Management
Test data both as input and output is managed by these tools. The test environment can also be controlled.

Configuration Management
A very large sector of the tools area. They are not generally considered part of the Test Automation area. Indeed more efficient productivity of the development process. They are designed to support the Software Development Life Cycle. Typical functions include secure storage, change control and build development.

Incident Management
Reporting and analysis are generated by these tools. They may also have work-flow to track and control allocation, correction and re-testing.

Test Management Tools
Creation, management and control of test documentation. They may also offer project management facilities for testing: scheduling of tests, result logging and management of incidents..

Runtime Execution

Performance Test
Performance tools have two main facilities. Load generation and test transaction measurement. Load generation is completed by either driving the application by using its interface. Alternatively test drivers simulate loads generated by application on architecture. Detailed measurements are taken and passed to a log file. More Load Testing

Character based and GUI Capture Replay
The tools simulate user-entered keystrokes and capture screen responses for comparison. Mouse movements, clicks and keyboard inputs can all be generated. A programmable language is often used to generate test specifications. Very important in regression testing.