[PROJ] "Fallback" support for transformations

[Even Rouault]

> While that approach could work for map rendering, it's not without
> issues. E.g. a point layer dataset with some points outside the extent
> of the grid -- when viewed at small scales you'll see all points, then
> as you zoom in past a certain threshold you'll suddenly lose the
> outside points when the grid shift kicks in. That's definitely going
> to lead to user confusion. It also won't help with other
> transformation tasks like layer exports, because we don't know the
> intended scale for the output we'd always have to use the
> user-specified grid based operation (and result in more bug reports:
> "i can see the points in the canvas, but when I save the layer I only
> get some from the centre of my map in the saved file").

Hard/impossible to make everybody happy on that subject. I would probably 
suggest going to your 2. behaviour (always return something, which was pretty 
much the behaviour with PROJ < 6), being the default. With an advanced setting 
for advanced users that want maximum geodetic pedantism to go to mode 1.
The cases where the grid based transformation are - generally (see below note) 
- in situations where points are outside of the area of use of the CRS (or are 
of use of the transformation), so you don't care a lot about accuracy of the 
transformation for those, because it is essentially unknown.

Side note:
There are a few cases also where the precise selection of a given transform / 
grid is not the ideal choice. For example (can't really think about another 
one) for the NAD83 -> NAD83(HARN) transformation, the NADCON4 grids used by 
PROJ are per US state. So, with selection of a single transformation, if you 
want to transform a dataset covering more than one US state, you're stuck. 
That's a annoyance that the data producer realized, and NADCON5 grids are for 
the whole conterminous US (but those are not integrated in PROJ for now)

Even

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