There are two kinds of execution-based testing:

• Testing to spec (black-box testing). Testers assume no knowledge of internals of the software. They test functionality purely based on the requirements and specifications, i.e. intended behavior. The software is given expected input, and results are analyzed for errors.
• Testing to code (white-box testing). This type of testing requires close interaction of our developers and QA engineers. Employs knowledge of the internal program logic to understand required input to exercise all execution paths in the software.

Testing to code is assisted by code coverage tools, such as Rational Pure Coverage. These tools provide information on what parts of source code were executed and how extensively. Code coverage is also indispensable as an integral part of test results, and in assessing readiness of software modules.

Incremental development not only means incremental integration of software modules, but also implies incremental or regression testing. Regression testing is re-execution of the previous test cases with the same data on every new release of the software. In simple terms it means: what was working in the previous release should NOT become broken in the new release.

There are always many builds throughout the project lifecycle. Thus automation of regression testing becomes an important task. Automation testing achieves:

• Increased productivity
• Reduced omissions, and repeatedly introduced software faults
• Shorter development cycle

Automation testing is implemented using various tools such as Rational Robot, Silk Test for GUI and Web-based interfaces, and programmatic tools and scripts for command line, and internal modules, such as database layers, business logic, and 'server-side'.

<• back