[PROJ] Proj 6 API questions

[Even Rouault]

> Done in https://github.com/OSGeo/proj.4/pull/1337
> 
> Now -- my final question on this topic (I hope!)
> 
> The interplanetary ellipsoid definitions in proj vary from those in
> QGIS' database. E.g.
> 
> Adrastea2000
> Proj: +R=8200
> QGIS: +a=10000 +b=7000
> 
> Thebe2000
> Proj: +R=49300
> QGIS: +a=58000 +b=42000
> 
> The QGIS ones were introduced by
> https://github.com/qgis/QGIS/commit/6055a4f566c16e4ae4686d5f666fa12408f2de03
> - but there's no source listed there, and I cannot find where these values
> were taken from, or why they differ from Proj.

Those are non-Earth ellipsoids. For PROJ, they have been imported from ESRI dictionaries at 
https://github.com/Esri/projection-engine-db-doc/blob/master/csv/pe_list_spheroid.csv
which are sourced from IAU 2000 apparently

The QGIS values seem conformant with
https://svn.osgeo.org/metacrs/sr.org/srsbrowser/data/IAU2000.wkt

So it seems ESRI tooks a spherical view, whereas IAU2000.wkt/QGIS an ellipsoidal one.

Interestingly, I just found
https://raw.githubusercontent.com/USGS-Astrogeology/GDAL_scripts/master/OGC_IAU2000_WKT_v2/IAU2000_v2.wkt
which uses the spherical view too.
I guess the _v2 implies some updates where done compared to the original
https://svn.osgeo.org/metacrs/sr.org/srsbrowser/data/IAU2000.wkt

CC'ing Trent Hare who might have some clues about where the "truth" is :-)

Even

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