How to Write a Test Plan for Software Development

Testing requires good planning, rigor and validity.

Testing is one of the most important activities of software development. It validates the purpose, process, and performance of the new software. Testing should not be an afterthought. It should be considered throughout the life cycle of the product. There are three levels of testing for software development: unit (individual modules), program (all modules) and system (integration of modules into the environment). According to Whitten and Bentley, in Systems Analysis and Design Methods, the approach and care taken to perform these steps in testing will be reflected in the quality and acceptance of the software product.

Things You'll Need

• Testing software

Instructions

1. Acquire a copy of the approved business requirements. Use this document to validate the intended purpose of the new system, detailing the business and functional objectives, including performance.

2. Align each objective to a specific testable measure. Never assume that an expectation will be satisfied by the completion of other components; define specific test scripts for each activity to ensure mutual agreement on compliance and performance. Publish these findings in the test plan and cross reference them to the business requirements document.

3. Develop test scripts for each level of testing. Maintain separate test scripts for the unit, program and system testing. Keep them updated as changes occur in the business needs and code design and use them as the product progresses through the implementation process and for integration (or restoration) activities in the event of a system outage or future implementation.

4. Acquire testing software. Testing software can assist in documenting test scripts, maintaining linkage to business requirements, capturing performance of ongoing and pending test activities, and reporting defects, discrepancies and unmet criteria.

5. Remember to perform positive and negative testing. Creation of new functionality is important. The system should know how to respond to data that is negative (or incorrect), null (missing) and anomalous (irregular) to ensure appropriate or acceptable responses in error handling.

References

• "Systems Analysis and Design Methods"; Jeffrey Whitten and Lonnie Bentley; 2007