[GRASS-dev] nasty gotcha in alternate ogr2ogr solution comment

[Paul Kelly]

On Wed, 5 Nov 2008, Patton, Eric wrote:

>>> AFAICT `g.proj -wf` doesn't include grid file, while `g.proj -jf` does.
>>> So I think my "best use `g.proj -wf` in that case" might be wrong as it
>>> drops the grid file, so really best to use:
>>>
>>> IN_PROJ="`g.proj -jf` +wktext"
>>>
>>> http://trac.osgeo.org/gdal/ticket/2638#comment:6
>>> http://trac.osgeo.org/gdal/changeset/15681
>
>> Mhh, that's fairly bad. Perhaps we need a note on
>> the g.proj and g.region manual pages?
>>
>> However. I feel that 'g.proj -jf' is a poor representation
>> compared to 'g.proj -wf', at least it doesn't help to
>> maintain metadata in a reasonable way.
>>
>> Markus
>
> Just so I understand the problem: g.proj fails to report nadgrids information
> with the -wf whereas -jf does? Should I go ahead and update the page, or is this
> a relatively simple fix in the code?

Well, there is no way of specifying a nadgrids file in the WKT format. 
Never has been - nothing new there as far as I know. What is new (to me 
anyway) is the +wktext PROJ.4 parameter - it looks to me like this should 
be included in the g.proj -j output by default - anybody disagree?

Perhaps also when GRASS is generating WKT in the GDAL-compatible format 
(i.e. -e flag not specified) it should also include this new 
EXTENSION["PROJ4"... parameter that seems to have been introduced - 
anybody any thoughts on this? It seems to me this might solve Markus's 
original problem. The whole thing seems like a huge hack to me but it's 
not our fault the WKT format is so deficient and anyway it's generally 
best to retain compatibility with GDAL as much as possible.

Paul