I like the visual method of writing scripts, but having used Selenium + SauceLabs.com on a real app and dealing with functional browser tests every day, I'm concerned about:
* Timing issues; most of the time spent writing functional tests is knowing when to wait for the next click (e.g. waiting for AJAX response and change in a certain DOM element)... and doing this in a clean way (e.g. WebDriverWait conditions instead of putting Sleep(1000) everywhere).
* Ability to expand the tests to do other things; for example, Selenium can take screenshots of the browser window because it works at a lower level. We use this for regression testing our web pages.
* Timing issues; most of the time spent writing functional tests is knowing when to wait for the next click (e.g. waiting for AJAX response and change in a certain DOM element)... and doing this in a clean way (e.g. WebDriverWait conditions instead of putting Sleep(1000) everywhere).
* Ability to expand the tests to do other things; for example, Selenium can take screenshots of the browser window because it works at a lower level. We use this for regression testing our web pages.