Performing User Acceptance Testing
Each item on the test plan (explained in the Test Planning page) should be performed
individually and carefully. If you think of more tests to try as you go, simply
add them to the test plan for later. Don't try to perform any ad-hoc testing amongst
the structured tests otherwise the test plan becomes irrelevant.
If a test succeeds and the results are as expected, simply complete the relevant
row in the test plan and move to the next test.
If a test results in error(s) or an incorrect functional result, create an entry
in the issue register. Describe in as much detail as possible (and practical) the
way the error occurred and how to get it to happen at will. Correction of a bug
that cannot be reproduced is often near impossible – because the fault can't be
pinpointed – so the steps to reproduce the error are critical.
Sometimes during the testing you might realise that something functions correctly
(according to the specification) but isn't quite what is now needed. Perhaps you
also realise that there is additional functionality that would be a great enhancement
to the software. This can also be recorded on the issue register. Obviously now
the focus of the issue register entry is not "What went wrong" rather it is now
"What needs to happen and why".
You may hit an error which prohibits you from accessing additional functionality,
in which case an error report with a high level of urgency is indicated!
Reporting Test Results
When all of the tests have been performed and the results recorded, all of the documents
are sent to Infosphere for review and the in-scope issues are addressed. Out of
scope issues (those not included in the original agreement for the software) are
new matters to be discussed separately.
|