quantum field theory - Distance and time measurement in the famous Superluminal Neutrinos Experiment

I tried to understand the technical aspects of the OPERA/CERN experiment, but apparently it takes some professional experience. Therefore I would like to ask someone better acquainted with such experiments to give some details concerning the setup in plain English.

The paper says they conducted a "high-accuracy geodesy campaign that allowed measuring the 730 km CNGS baseline with a precision of 20 cm". They also describe the clock setting and synchronization, but I'm not sure I understand (there is something about "master clock"). So:

1. 
   

   I figure that for some reason it wasn't possible - in order to obtain better accuracy than 20 cm - to run photons first along the track, and calculate the distance based on the velocity $c$ and time measured (or simply compare the times for photons and neutrinos), right? Was it because such "contraptions" do not allow for measuring photons or for some other reason? EDIT: I do realize now it's (practically) impossible.

   
   



2. 
   

   How was the time measured? Were there two clocks at the beginning and at the end of the track, or was there just one clock hooked up (with the wretched fiber-optics cable) to both ends? If there were two clocks, were they synchronized through a GPS satellite or some other external device, or was there some sort of direct synchronization procedure? (What was it?)

   
   

To make sure: I looked at the paper before asking, but I am not really sure what they mean. I am not a professional particle physicist, and I have no knowledge whatsoever concerning the devices used, as well as procedures, technical jargon, etc. That's why I asked for help, and will be much obliged for information given in "plain English".

EDIT: In his answer dmckee said that: "Both the timing and the distance were established with the help of GPS." Does it mean that the margin of error is the same for both measurements - time and distance? If the the distance (of 730 km) was measured with 20 cm accuracy, would there be a margin of error for time (clock synchronization) of the same proportion? After all, time is distance and distance is time.