Nice rebuttal and a lot to discuss here :) I'd like to concede 1 point in that I don't know enough about Robot framework and that a lot of my argument is based on other implementations I've seen with Selenium and QTP and implementations I've done myself in the past. Can you please share your production level test code so that I can better understand what you feel makes KDT a good option. No theoretical examples please. I'd like to see and understand actual code testing real software.