Sorry, have to disagree about the icewater method. When doing a one point cal (which is by far the most common), you want to calibrate at the temperature nearest to the temperature that you will be measuring, which is anywhere from 140 degrees to 210 degrees. Water boils at around 210, so it is perfect.
The problem with a one point cal is you can have drift, and while you may be rock solid accurate at 32 degrees, by the time you get up to 200 degrees you may have drifted quite far off target. Not all temperature gauges support a two point cal (I don't even know of one that does, but I'm sure they are out there). So if you're stuck with a one point cal, the higher temp makes much more sense to me. I could be wrong, maybe someone with a physics background can chime in here.