[CF-metadata] URNs

From: Jonathan Gregory <j.m.gregory>
Date: Wed, 3 Oct 2007 17:44:52 +0100

Dear John and Bryan

OK, I accept the argument that the URN could give access to *additional*
information which is potentially useful. If we include it on that basis, it
is a by-product that for applications that understand the URN it provides an
alternative to the attributes that describe the CRS. If I were writing a
generic application to plot data from CF-netCDF files, however, I would need
the parameters that describe the projection explicitly. (I can't give a URN
to the IDL map_set procedure, for instance.) If I want the program to work
for any projection, regardless of whether it has a URN or whether the URN
has been recorded in the netCDF file, than I need to be able to interpret the
projection attributes themselves, and not rely on looking them up elsewhere
using the URN. So if I'm going to do that coding anyway, the URN doesn't help
in the end with the task. Its value would be in other descriptive information
it might give access to, such as John describes.

Cheers

Jonathan
Received on Wed Oct 03 2007 - 10:44:52 BST