On the fractional Harry Dym equation