[GRASSLIST:5055] Re: [GRASS5] Re: Re: rasters reprojection UTM -> LAEA

[Paul Kelly]

On Tue, 23 Nov 2004, Glynn Clements wrote:

>
> [CC'd to developers' list]
>
> Rado Bonk wrote:
>
>> GRASS:~> g.setproj:
>>
>> XY-location cannot be projected.
>
> In that case, you'll have to edit the cellhd file manually to change
> the projection to something else.
>
> Question for developers: why is g.setproj paying any attention to the
> current projection type?

I suspect it may be a historical artefact from when GRASS supported only
XY, UTM and LL and g.setproj was a 3rd-party module added by the Soil 
Conservation Service (who originally wrote v.proj and introduced PROJ into 
GRASS via Mapgen). The location would have already had its projection 
defined in the WIND file and g.setproj's original job was just to augment 
this by adding PROJ_INFO and PROJ_UNITS files (for use by the pj_* 
functions), not to change the projection of the location to something 
else.

So I think this original design philosophy of augmenting a location's 
projection description rather than overwriting it with a new one may 
explain the behaviour. But g.setproj is in such a mess I wouldn't like to 
have to change it. In 5.7 g.proj can be used to create a location (or 
overwrite the projection of an existing location) if you 
have a PROJ.4 or WKT projection description, an EPSG projection code or a 
georeferenced raster or vector file. Some day hopefully g.setproj can be 
done away with (when someone in the PROJ community produces a better 
database of the parameters required by each projection than is currently 
contained in lib/gis/geo_init.c).

Paul