[GRASS-dev] Improved ellipsoid name support

Yesterday I committed to CVS some updates to the gproj library that I'd had sitting around for ages. They do the following things:

1. When generating a WKT projection description, put the full ellipsoid name from ellipse.table in the "SPHEROID" section. The short name stil goes in the "GEOGCS" section.

2. When generating a GRASS PROJ_INFO file from an input projection description, try to put the ellipsoid name into the PROJ_INFO where possible, instead of just a and es parameters. People had complained about this before. The new functionality works as follows:
  a) If a datum name is specified, use the ellipsoid associated with that datum (i.e. ignore the ellipsoid parameters)
  b) If there is no datum specified, try and match the ellipsoid parameters against the known ellipsoids in ellipse.table (within a specified tolerance - see the code in lib/proj/convert.c around line 521; comments welcome) and use the first one that matches
  c) Only if a) and b) both fail, put the a and es parameters into PROJ_INFO directly as present

These changes should make the PROJ_INFO files generated by g.proj (and r.in.gdal, v.in.ogr) more like those generated by g.setproj and should hopefully make it more viable to use these functions as some kind of total replacement for g.setproj in the future (at present I'm not sure how!).

Paul

Paul Kelly wrote:

1. When generating a WKT projection description, put the full
ellipsoid name from ellipse.table in the "SPHEROID" section. The
short name still goes in the "GEOGCS" section.

I'd like to update the "GIS concepts" part of the GRASS Wiki.
  http://grass.gdf-hannover.de/wiki/Gis_Concepts

WRT proj jargon, my understanding is:

ellipsoid and spheroid are the same thing, some people use one, some the
other.. and ellipsoid is arguably the better choice ?

datum = [ellipsoid, origin] ?

thanks,
Hamish