[Freegis-list] Newbie Question

[Markus Neteler]

On Mon, Feb 27, 2006 at 04:52:32PM +0200, Ari Jolma wrote:
> Markus Neteler kirjoitti:
> > Sounds great. There is one thing which I discussed with Howard
> > Butler at Chicago recently: The GRASS code should not be
> > replicated in the SWIG interface. This will require changes
> > in the GRASS code, in a sense, that main procedures (parameter
> > definitions etc) are separated from the real algorithm.
> > This is unfortunately not the case for the example (r.slope.aspect).
> > 
> > I am happy to get suggestions how to change the code in GRASS
> > to be easier/better wrapped with SWIG.
> 
> 
> I'm hoping that it is possible to write a translator, which automates
> this step, i.e., the example r_slope_aspect.c would be generated with
> that translator. If it needs a full C parser, then it gets complicated
> however. It's a bit similar to what Swig itself does when it creates the
> wrapper C code.
> 
> 
> > 
> >> Can Grass use geodata via GDAL? Or is it only the other way around,
> >> i.e., GDAL can open Grass data?
> > 
> > Both, I think:
> > - GRASS imports external data via GDAL/OGR (r.in.gdal/v.in.ogr)
> > - GDAL/OGR can read data via GDAL-OGR-GRASS plugin
> > 
> > So it may be an idea to wrap r.in.gdal and/or v.in.ogr into
> > GRASS-SWIG. The someone could remotely import data (even
> > create a new "location" which is the GRASS project) and work
> > on these data from outside via SWIG. Results can be read
> > directly via GDAL-OGR-GRASS plugin, e.g. in Mapserver.
> > 
> > Like this GRASS operates as GIS analysis backbone.
> 
> 
> The ultimate tool would also have R spatial capabilities among other
> things... The problem is how to move the data from one environment to
> the other. Probably that could be organized through GDAL/OGR, i.e., all
> tools may have their native formats but should include GDAL/OGR i/o.

I always wanted to look at the R to C interface (embedded R or however
that's called to launch R tasks from the GRASS side).