Distance vor antipodal points

[Anonymous]

Hi,

I had a look at your GeodSolve tool (just to check the result of your lib before implementing it into my project ;) )

I tested the inverse problem with following data:

Equitorial antipodes:

ellipsoid (a f)     = 6378137 1/298.257223563 (WGS84)
status              = OK

lat1 lon1 fazi1 (°) = 0.000000000 0.000000000 0.000000000
lat2 lon2 fazi2 (°) = 0.000000000 180.000000000 180.000000000
s12 (m)             = 20003931.4586

Polar antipodes:

ellipsoid (a f)     = 6378137 1/298.257223563 (WGS84)
status              = OK

lat1 lon1 fazi1 (°) = 90.000000000 0.000000000 180.000000000
lat2 lon2 fazi2 (°) = -90.000000000 0.000000000 180.000000000
s12 (m)             = 20003931.4586

Own antipodes:

ellipsoid (a f)     = 6378137 1/298.257223563 (WGS84)
status              = OK

lat1 lon1 fazi1 (°) = 35.784500000 -5.808666667 -0.070197165
lat2 lon2 fazi2 (°) = -35.784500000 174.191933333 -179.929802835
s12 (m)             = 20003931.4254

For me the results are very surprising because I thought the geodesic, which connects the poles must be shorter than an antipodal geodesic on the equator. However, all my results have the same distance.

Am I wrong? Is there a bug? Did I miss something?

Any explanation is appreciated!

Thank you in advance,
Mark

[Charles Karney]

The shortest path between two antipodal points on the equator is via one of the poles (not along the equator).

[Anonymous]

Ah... of course, you're right. I missed this. Thank you!