I think you are right that
the Distance function assumes the earth is a perfect sphere, and there
is probably a formula for correcting the results of Chris' lat/lon
extraction code. However, when we tested the accuracy of the lat/lon
values calculated (using MapInfo as the referee), the deviations were
very minimal. Anything within a 1,000 mile range from the point the
Distance was calculated from appeared to be within feet of the true
location. Maybe the accuracy deteriorates the closer you get to the
poles, but I guess that is not such a big issue for most of us.