Distinguish verification, validation, quality assurance and testing using evidence from one case.
Write a release claim whose scope does not exceed the available evidence.
When a team says a booking system has ‘passed testing’, ask which behavior was tested, with which data, in which environment, and against which expected result. Testing gathers information about a product through observation and comparison; it cannot establish that no defect remains. Verification asks whether an item was built according to its specification. Validation asks whether the built item serves the intended user need. Verification can pass while validation fails if the specification omits a real need. Quality assurance has a wider process focus: organizing work to prevent and detect problems, including requirements reviews and learning from failures. These terms answer different questions rather than competing for a single label.
Build an evidence chain before broadening the claim. A defensible narrow statement is: ‘One request in the test environment returned success and a message matching the submitted time.’ Attach the input, build version, server and user time-zone settings, and the way the stored value was independently checked. If storage was not inspected, a correct-looking message is no evidence that the database value is correct; the interface could echo the input while saving another value. If cross-zone use was not tested, the traveler experience is still unknown. This discipline helps a decision maker request the missing evidence instead of accepting a vague ‘everything is green’ status. A stopping decision must consider the risks that remain unobserved as well as the number of checks run.
Use three steps before issuing a quality judgment. First define what is required: the written rule, the user need and any unresolved assumption. Second design a repeatable observation: known input, known initial state, expected result and an independent way to inspect it. Third compare the observed result and state the scope of the conclusion. If the message is right but the stored record is wrong, comparing storage with the rule reveals a defect. If both are technically right but a user cannot identify the time zone, the gap concerns fitness for use. The appropriate change differs: repair the code in the first case; improve the requirement or interface and add a new check in the second. That distinction carries more information than marking every run simply ‘passed’.
Lesson summary
Distinguish verification, validation, quality assurance and testing using evidence from one case. Write a release claim whose scope does not exceed the available evidence.
