The second detail is conceptual: keys pressed by a screen reader users are not necessarily directed at any particular element, so the WebDriver command's required element reference would be superfluous in many cases.
Just a small nit, the Perform Actions command can be used to generically inject input to the page that's not targeted at a specific element. I sometimes use this for testing keyboard accessibility. Not that it changes your conclusion.
https://github.com/bocoup/aria-at-automation#press-keyboard-keys
Just a small nit, the Perform Actions command can be used to generically inject input to the page that's not targeted at a specific element. I sometimes use this for testing keyboard accessibility. Not that it changes your conclusion.