[gdal-dev] geolocation info in netCDF: "northernmost_northing" etc. not universal

[Bruce Raup]

Frank,

Thanks for the reply.  I mistakenly said GeoRef when I meant to say
GeoTransform.  Your description of what would happen if we make the
change confirms what we thought.

I think we will go ahead and make the change, and file a ticket.  Lots
of data are in polar projections, and it seems like a small amount of
work now could save a lot of confusion in the future.

Best regards,
Bruce


On Wed, Jan 27, 2010 at 17:44, Frank Warmerdam <warmerdam at pobox.com> wrote:
> On Wed, Jan 27, 2010 at 2:22 AM, Bruce Raup <bruceraup at gmail.com> wrote:
>
>> We propose using terms such as
>>
>> grid_top_boundary_projected_y
>> grid_bottom_boundary_projected_y
>> grid_left_boundary_projected_x
>> grid_right_boundary_projected_x
>>
>> Alternatively, grid corner coordinates could be used.
>>
>> What would the impacts of such a change of names in our files be?  If
>> the GeoRef field is also in the file, would it break existing software
>> for being able to navigate (geolocate) the data?  And, would the
>> community consider the adoption of more universal terms more generally
>> (i.e. in GDAL itself)?
>
> Bruce,
>
> I am not really an expert on the GDAL netcdf driver, but
> I will attempt to respond.
>
> First, I think it would be better to just document that the
> existing tag names (eg. northernmost_northing) should
> just be interpreted with the new definitions
> (grid_top_boundary_projected_y) rather than trying
> to change the actual tag names in order to gain some
> possibly clarity in less common situations.
>
> I have reviewed the GDAL netcdf driver and it seems to
> look for two things.  First a GeoTransform value (which I
> think is only a GDAL convention).  If that is not found it
> looks for northernmost_northing, etc.  I presume these are
> part of the CF conventions, though I don't know for sure.
>
> I do no see any reference to a "GeoRef" tag.
>
> So if you make the planned change, and if you do not
> also provide the GeoTransform then older versions of
> GDAL would not be able to read the georeferencing of
> the files.  I presume this might also be the case of
> almost all other software, even if you did support
> GeoTransform.
>
> However, if you do go ahead with this convention, then
> please file  ticket with a small sample file so we can add
> support for it.
>
> Best regards,
> --
> ---------------------------------------+--------------------------------------
> I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
> light and sound - activate the windows | http://pobox.com/~warmerdam
> and watch the world go round - Rush    | Geospatial Programmer for Rent
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/gdal-dev
>



-- 
Bruce Raup
http://cires.colorado.edu/~braup/