I have implimented Vincenty's Direct and Inverse algorithms with both VB6
and Delphi7. I have been distributing the code as open source.
A user has pointed out a problem in the Inverse function which I had not
noticed. The problem exists in both the VB6 and Delphi versions. I have
been trying for several days to fix it. The essence of the problem is:
given that point 1 is at a fixed latitude/longitude (I have been using
89N/0E) and compute a the length of a geodesic at say 20N/0E the length is
1,572,913.307 meters and when doing the same for 20S/0E (trans-equatorial)
the length is the same. Obviously this is not correct. I believe the length
of 89N/0 to 20N/0 is correct. I am using WGS84.
My algorithm is taken directly from Vincenty's paper as published in "Survey
Review" XXII, 176, April 1975 entitled "Direct and Inverse Solutions of
Geodesics on the Ellipsoid with Application of Nested Equations"
I would like to correspond with anyone that might shed some light on the
problem. I am confident that the equations faithfully follow those in the
paper, because they yield good results for known test cases of data that do
not violate the conditions set forth above.
Any enlightenment will be appreciated.
TIA
Vic
--
__________________________________________________ ______
Victor Fraenckel - The Windman
vfraenc1@nycap.rr.com
KC2GUI
www.windsway.com
Home of the WindReader Electronic Theodolite
Read the WIND
"Victory at all costs, victory in spite of all terror, victory however long
and hard the road may be; for without victory there is no survival."
- Winston [Leonard Spencer] Churchill (1874 - 1965)
Dost thou not know, my son, with how little wisdom the world is governed?
-Count Oxenstierna (ca 1620) to the young King Gustavus Adolphus