I am almost done. While working on the regression test I found a significant
bug in the original earth distance package, so this really does need to
get updated. While I was doing that I switched it to use the haversine
formula as that is more accurate for short distances than the formula
they used previously.
