sample application that broke test execution
There are TestComplete and HP QTP. I need window- and web-application under the test, so that test will fail either on TestComplete or QTP.
For example, complex object recognition will fail , causing test fail.
And more.... Could you please explain, why TestComplete and QTP writes test scripts and not test compiled code ?
I'm not clear on the first question, but to test across windows & web applications you can try Squish GUI Tester
Regarding scripts vs compiled code: Think portability and how quickly you can update and re-execute tests if you don't have to compile them. They can also remain more platform independent. Not sure which of those apply to TestComplete and QTP, but the cross platform nature of scripting languages instead of compiled code definitely applies as an advantage when using cross platform gui testing tools like Squish.
wiki article which *may* help: Interpreted language - Wikipedia, the free encyclopedia
On the 2nd question...
From a technical perspective, I believe both leverage WSH ( Windows Script Host - Wikipedia, the free encyclopedia ) technology to avoid having to implement their own scripting engine or compilation.
As from a usability perspective. Scripting is easier to debug. Does not have the performance profile as compiled code, but during debugging you can pause at a break point, and in the interpreter window evaluate statements, shim in code you want to test, perform manual intervention and experiment while stepping through the debugger.