Sorry Mac, I am not aware of a HTML version of the article, although I would think there should be one. I believe it appeared in the March 2003 Issue, Physics Today.org. It is very slow to open on my computer, but I did once last night and again today. When I say slow, click the link, go make yourself a sandwich and eat it, ha! Then come back and check. Really, there are some graphics on the link and pdf is always slow on my connection anyway. One I have been trying to access has crashed my computer twice. This one worked, it was just sloooow. As far as the GPS system is concerned, Mac I keep saying, there is no way to get accurate figures for relativity effects. The prelaunch offset is just a prelude, a simple artifical frame of reference (ECI) is used as a baseline to synchronize clocks, as they cannot be synchronized in a rotating frame. But a rotating frame is also used and that is where most of the relativist effects are accounted for before the code is sent by the satellite to the ground reciever. Each individual sat clock has its own frequency changes and those changes even differ in one orbit of the satellite, based on modelling. Even the Earth's difference in gravitational effects at various points is incorperated into the satellite's signal it broadcast. Some recievers for fast moving objects such as jet aircraft even take the first-order Doppler shift into account, in addition to the second-order Doppler shift due to satellite velocity relative to an Earth-based reciever. The prelaunch offset is just the bare beginnings of the frequency adjustments that are made to the clocks, a common reference point.