Inspired by my work at JPL, I wrote a small utility that allows you to compare distances between two points using the two most common methods:
The basic idea is that the great circle distance (compared to regular Euclidean distance) method assumes that the Earth is a perfect sphere, but Vincenty's formulae more accurately approximates the Earth as an oblate spheroid.
To see this effect in action, notice the error measurements for distances computed near the equator and those near the poles; for example, compare the distance between the Natal, Brazil (-5.733987, -35.209868) and Abidjan, Cote d'Ivore (5.415605, -4.022813) -- the percent error between Vincenty and GC is 0.081%. However, when comparing say, Saint Lewis, Canada (52.371317, -55.684063) to Alta, Norway (69.972182, 23.270742) -- the percent error is larger here, at 0.122%.
Per Wikipedia, the Great Circle distance is roughly accurate to around 0.5% of the true value, whereas Vincety's method is accurate to 0.5 millimeters, anywhere on the Earth's surface.
Enter comma-separated coordinates in the form:
(<latitude>, <longitude>)
