Why is ice slippery?

There's an article in the December 2005 episode of Physics Today written by Robert Rosenberg, titled 'Why is ice slippery?'
A simple question, but the answer is however not so simple. The first part of the solution is the following: for example, when you're on ice skates the pressure (force per unit area) is very high (the blades of the skates occupy a small area, whereas the weight on them is large). This increase in pressure results in a lower melting temperature of the ice below the skates: -3.5 degrees Celsius. Thus, the ice melts, and you skate over the water. Simple. Or not?
However, what happens when the outside temperature is below -3.5 degrees? Can't you skate below that temperature? Nonsense. Also, the pressure on the snow exerted by a skier is not enough to make it melt (larger area). What happens when you slide over the snow or ice, is frictional heating. The author then mentions an experiment, with the following brilliant remark:

"The increase in temperature with velocity, they observed, was consistent with frictional, localized heating of the ice underfoot to create a thin water layer. Were pressure melting -an endothermic process- the dominant contribution, the researchers would have expected a decrease in temperature."

The heat created due to friction causes the temperature to rise, so a waterlayer is created, and you can skate over the ice.
There, problem solved. Not quite.
But why can ice be slippery when you're standing still on it? This question is analyzed in the remaining pages, where he shows what evidence there is for the existence of a liquid-like film at the surface, even at temperatures below zero. He concludes with a discussion of experiments about the thickness of the film and temperature-range of the effect.