Earlier quoted context omitted.
And that is where the vibrate comes in. the keyboard is of a known layout, and by marking by feel when you move from one key to the next you know where you are. This allows the keyboard to be operated without looking, much like how most people use a ordinary keyboard today once they are past the hunt-and-peck stage. A big reason for touch screens being a problem in many situations is that you have to actually look at…
I'm guessing you downvoted me because you disagree with my point of view. I'll just tell you this: I have been thinking a lot about the stuff you talk about here and even prototyped a hardware device to try out the idea. I had the same point of view (It would be nice to have an input device I can use without even looking) but after experimenting I realized this isn't the case. Also this "vibration" thing you're talki…
Combine with pressure detection and you can make Swype / SwiftKey provide a sense of where you're dragging your finger when swipe typing, plus that you can put emphasis on the right letters easily when swiping across groups of letters that would make for different words, making for better precision in the word suggestions.
It would also be much easier to implement gesture support by having areas much easier to target that you start gestures from or end them in. You'd also instantly get tactile feedback telling you if you made the gesture right or not, like for switching character sets or for moving the cursor.
Lots of gestures could be supported, where you'd essentially be "asked" with tactile feedback to push harder after making the gesture to activate the particular action. Suddenly you solved both accessibility, discoverability and precision, while reducing the error rate.