I'm philosophisizing on some GPS ideas and have not played with these things before. I hope someone has some knowledge or where to find knowledge knowledge.

I was wondering and have google searched w/o success about the following

If I read the GPS location of point 'a' with whatever error single point gives me, how long do I have to get to point 'b' to get a fairly precise reading on distance and orientation between the two (the error should be the same and thus cancel).
This assumes something I do not know for sure: "when reading two points within 100m of each other by GPS the error for both will be the same"

What would be wondeful would be a curve for the above: error (distance and orientation between the two) vs time between the readings.

Also will there be a jump in the reading when a "satelite switch" happens

Erik