Why bother? I have found the UI is normally reasonably stable. If it is broken the application tests will often find it.
It is used by so many for so long that specific test for open, close, resize, box sizes, corporate colours, logos, et al will be observed pretty quickly without any automated validation
K1W1, that is a revelation: you need to have a stable UI interface before a testing tool can use that interface to test the application behind it. If it is broken the test tool cannot properly use it, and it will be unable to perform any tests.
QAVizee, your question is stated incorrectly. If you have a GUI interterface you need one set of libraries, whereas if you use a character UI the most basic interface routines will be used. Most test tools work the same kind of magic, but not all tools work with all interfaces. So, in order to give examples of tools you need to tell us the interface(s) you plan to test with first.
Exactly my point.
If the new release has broken the UI why would I run a UI test when the app test will find the problem equally as well?
If the UI is stable, running the UI test and the app test will both pass.
Why would you bother with a UI Test.
And your point is?
You agree that we do not need to bother with a specific automated UI test as there are normally sufficient other test to reveal UI problems?
Or are you saying a smoke test is a de facto UI test?
The monkey test should also find broken UI.