I agree. Voice control in the Apple/Android paradigm is... weird. Talking to the phone, requiring Internet connection, unusable on a busy street. Take out the phone, unlock it, move it in front of your face (now I can't see a thing and start bumping into people), "ok Google", say it again, wait for the S4 to react, say something, see it's interpreted wrong...

I very much prefer to type, it's faster even on a touch screen.

So here I am, waiting for a subvocal recognition system and the comeback of off-line speech processing.