I don’t think it’s poor user interface at all. When you see “Press S” on the screen, you should understand it as “Press the key that produces S”. If you chose to change the mapping, that’s your responsibility to find that key.

Also, the whole DVORAK superiority thing is complete myth…
http://www.reason.com/news/show/29944.html

If I have that wrong I apologize; but it seems like a poor user interface. I think a better approach (assuming you aren’t going to provide full key-mapping functionality, which may be out of scope) would be to convert the QWERTY keys you are interested in at the beginning to the keymappings the user has in place, and then use those mappings to display messages (and look for those keypresses). E.g., modify the string “Press %c to continue” based on what character is generated when the user presses the key two to the right of the CapsLock key.

By the way, might be a good idea to mention how you assign it in the article, since it’s different from KeyboardState:

LocalizedKeyboardState ks = new LocalizedKeyboardState(Keyboard.GetState());