I'm not sure I understand. To drag select text on android I just long press on text and drag my finger. It leaves little control points on each side to adjust to the character level (otherwise generally snaps to whole words). The first two steps seem unnecessary - why not just press on the word to start with?
Because pressing on words involves taking your finger off the keyboard, and the text is anyways often too small to be precise. You can fix it with the control bars on the sides, but it's more finicky and more interaction and requires you to move your hands.
Perhaps I misunderstood though - is this hard press when the cursor is on the word.
I guess I just don't really see the problem it solves, it's just moving my thumb over the text I want rather than three distinct actions just to start the selection. Then I guess two more to change the selection and end the selection process? Whereas to select text I just long press -> move. Or for one word, long press.
Precision seems fine as I test it here, and if I want sub-word selections (rare) then adjustments seem like a reasonable approach. If the selection on iOS snaps to words anyway (second force press selects whole word?) you'd need this on both platforms.
I don't really understand how the android approach requires more interaction. Press & drag is surely less interaction than press (get cursor), drag (position cursor), press (select word), drag (move cursor), press (finish selection).
The only selections I generally want without moving "off" the keyboard is deletions, and that's accomplished just by dragging from the backspace.
