Nevertheless in the 8 years I have been flying behind the old KLN94, I have never seen it display an altitude more than about 30ft (i.e. 1 millibar) different from the published airfield elevation.
That's a pretty good record, even before one starts to compare it against mis-setting the QNH by 10 points (300ft), flying 1000ft too high/low (due to misreading the analog altimeter) etc.
GPS altitude is a lot better than most people think - unless the GPS is a piece of junk from a camping shop, and even the modern ones of those can be very good. My Garmin 496 (OK, not from a camping shop) is within 5-10ft of the airfield elevation and most of the error is probably the height of the unit above the concrete
Of course it doesnt but you can get the wind from the tower.
If there is any real wind, say 10kt+ reported SFC wind, I would be more careful using a GPS for flying the final approach, since there is likely to be a significant wind shear in the last 100ft or so. But it should still be better than nothing.
When IO said there was no way of knowing if the calibration was correct, he had probably forgotten that you can ask for a mode c check from just about any radar unit every time you fly if you want to. On modern units you can also display the encoder output on the front panel and compare it with your altimeter set on 1013.
I didn't forget; I was referring to low end GA not carrying transponders (for all kinds of reasons; real, artificial, or Civil Liberties driven) but presumably still wishing to carry some form of TCAS. If you want to do TCAS on the cheap, nothing beats GPS (i.e. some form of ADS-B) for low lower, simplicity, and accuracy.
By way of experiment I have sat on the ground at a point of known elevation and compared a number of GPS receivers that I have, including a couple specifically intended for aviation use. The altitudes indicated have varied significantly
What models were these?