[CF-metadata] GIS versus CF conventions

[Magnus Hagdorn]

Hi all,
The georeferencing data and lat/lon grids are redundant. I guess, in the
case of observational data the grid might not be regular, on the other
hand model output will use grids which can be georeferenced. So why not
require at least the projection information or lat/lon grids. Certainly
in the case of staggered grids having two lat/lon grids which are
off-set by half the grid resolution is somewhat silly.
Cheers
magi