Thanks a lot - very clear explanation of a complicated subject.
Since my posting I got bits and pieces from the usegroups. A couple of interesting tidbits: almost all of the altitude measurements displayed in the 60CS's fields are readings from the altimeter. The only GPS altimeter field seems to be the one accessible from the satellite page (which I had not discovered before).
The logic of the altimeter/GPS combination is as follows. GPS elevation data is less accurate than your long/lat position. It is also much more subject to jumpiness because of the nature of GPS reception, changes in satellite geometry, etc.
The elevation data from the altimeter is generally more accurate than the GPS measurement: 1) if properly calibrated; and 2) assuming there is not much weather-related change (i.e. non-altitude connected change) in the barometeric pressure. It also gives much steadier readings than the GPS-derived figures.
My understanding of the interplay is that the primary driver of the altitude measurement is the barometer, and this is what is reported in the tracklog and I presume the datastream. However, the GPS-derived reading will be used to correct the barometric reading if there is a large disparity. This will happen if a weather front comes in quickly.
Before doing any calibration, I was consistently getting altitude readings of -100 feet at my home (real elevation +17 feet). I will experiment and see how it goes. It's fun learning about this stuff.
I wish the user could choose whether the altimeter is on or not. Sometimes rough numbers are enough, or you don't have a source for calibration.