 Run the Tests
Once the scripts have been written, they can be "run";
that is, followed step by step. Each test case should be marked
with the name of the tester, the date of the test, and whether
it passed or failed. The tester should sign the test case at
the bottom, as proof that it was run. A test log should be maintained
showing the result of each test; if you added a Pass/Fail column
to the Traceability Matrix, then that is where you record the
result (photocopy it first so you can use it again for subsequent
runs).
When a new software release is received, the tests must be
run again. The Test Manager may decide that not all of the tests
need to be run, just a subset, but this depends on the application,
the time available, and how much confidence you have in the product
being tested.
An automated test tool, such as Mercury WinRunner or IBM Rational
Functional Test, may help speed up this re-testing, which is
called "Regression Testing". However, if the screen
layouts are changing frequently, an automated tool is not of
much use. You can spend more time fixing the automatically recorded
scripts than actually testing. |
|