[gdal-dev] GDAL 3, PROJ 6.2 exportToProj4() drops Datum silently

[Even Rouault]

On jeudi 7 novembre 2019 15:24:44 CET Nyall Dawson wrote:
> On Wed, 30 Oct 2019 at 06:16, Even Rouault <even.rouault at spatialys.com> 
wrote:
> > > I realise this too, but to permit legacy results to be replicated,
> > 
> > Some legacy behaviour was just plain wrong/inaccurate. Trying to replicate
> > it in PROJ would be heartbreaking, an extra maintaince burden, and a
> > confusion for PROJ users because depending on the path they take, they
> > would get different results.
> > 
> > Not-so-random example with GDAL 2.4/PROJ 5 with GDA94 to GDA2020
> > 
> > You get a ~1.5 metre shift to the north east as expected.
> 
> This thread has been bouncing around in my mind for the last week.
> Forgive me if I'm misinterpreting the situation, but wouldn't you
> explicitly want legacy results NOT to be reproducible in cases where
> proj 6's more accurate transformation capabilities have affected these
> results?
> 
> If a particular study isn't affected by relatively small shifts such
> as the 1.5m shift in the GDA2020 situation, then the results should
> remain reproducible even if this shift is corrected. But if a
> particular study IS affected by distances of this magnitude then I'd
> argue that it would be vital that the older proj behavior can't be
> reproduced exactly, and accordingly the previous study/results are
> invalidated as a direct result..!

Are you saying that the new behaviour of ExportToProj4() shouldn't cause 
issues ? Well, that's more subtle than that. As +towgs84 terms might be 
omitted more often now, you can get differences of ~ 100 metres when dealing 
with datums dating before the GPS era. So when using GDAL 3 ExportToProj4() 
you can effectively get "wrong" results in some cases.

The action I took for now is to document the issue with ExportToProj4():
https://gdal.org/doxygen/
classOGRSpatialReference.html#a271b3de4caf844135b0c61e634860f2b


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