[PROJ] Vertical deflection: of interest for PROJ ?

[Even Rouault]

Hi,

While integrating different datasets over the last months, I've noticed that a 
few geoids (like the USA or Australian geoids) are accompanied with the angles 
for the vertical deflection. I've read a bit about that, but I'm still unclear 
if such data would be useful for what PROJ does, that is coordinate 
transformations.

For the formula converting between ellipsoidal height to orthometric height, 
H = h - N, sometimes instead of the = one sees a ~= . My understanding of the 
typical deflection angles is that the approximation is legitimate, as being 
largely a sub millimetric phenomenon, given that the cosinus of small angles 
is super close to 1.
But would there be an impact on horizontal coordinates ?
For example for a difference of height of 1000 m and a deflection of 50 arc 
second, that would be 1000 * tan(50'') = 0.24 m
But perhaps the use case for a PROJ-like context could be "given a skyscrapper 
1 km high whose facade is vertical along the plumb line, given the coordinates 
at the bottom, what are the coordinates at the top ?", which doesn't seem to 
be something people would use routinely :-)

Even

-- 
Spatialys - Geospatial professional services
http://www.spatialys.com