What if you had a OCR overlay that had a DOM like structure along with coordinates? You could hit a key combo, and then the text dom could float over the page and you could select nodes of it using your keyboard modifiers of choice (vim, emacs, etc).
You could also read the process memory directly by a supervisory program, like a debugger, to extract or insert text directly. It would be like idea implantation.
You could also read the process memory directly by a supervisory program, like a debugger, to extract or insert text directly. It would be like idea implantation.
Does that make sense?