Pan gesture: why need setTranslation to zero?

Question

I add a pan gesture to a view, move the view while finger moved, but I found if I do not call recognizer.setTranslation(CGPointZero, inView: self.view), translation is not right. why ?

  @IBAction func handlePan(recognizer:UIPanGestureRecognizer) {

    let translation = recognizer.translationInView(self.view)
    recognizer.view!.center = CGPoint(x:recognizer.view!.center.x + translation.x,
      y:recognizer.view!.center.y + translation.y)
    recognizer.setTranslation(CGPointZero, inView: self.view)// this line must need, why?
...
}

Answer 1

I don't speak English well, but I think it may be enough to explain this.

A translation in UIPanGestureRecognizer stands for a vector from where you started dragging to your current finger location, though the origin of this vector is {0, 0}. So all you need to determine the distance you dragged is another point of this vector. You get this point by calling :

recognizer.translationInView(self.view)

Then this point helped you setting a new location of your view. But UIPanGestureRecognizer is indeed a "continuous" reporter, she will not forget the state after the last report. she didn't know that you have used up that part of translation(to re-locate your view), so the next time when "handlePan" is called, the translation is not calculated from previous location of your finger , it is from the original place where started your finger dragging!!

That's why you have to call:

recognizer.setTranslation(CGPointZero, inView: self.view)

every-time you used that translation to re-locate your view, as if you are telling the recognizer that you are going to start a new drag gesture.