[CF-metadata] URNs

From: Jonathan Gregory <j.m.gregory>
Date: Thu, 4 Oct 2007 11:50:14 +0100

Dear Phil

Thank you for considering all these comments carefully.

> I don't think we should assume that a URN necessarily needs to be
> resolvable (referenceable?) across the internet in order for a software
> application to work with it. If an application 'understands' the meaning
> of a particular URN then all's well.

> It would be perfectly feasible for an organisation to maintain a local
> cache of URNs ...

It would be, certainly. It is also possible for applications used to analyse
netCDF files to be made aware of how to translate URNs. However, the practical
constraint is that they are *not* currently aware on the whole. It is right to
put in place mechanisms which will allow a better software environment to be
constructed, but looking at these issues from the point of view of a *user* of
data and software, my principal concerns are reliability and ease of use.
If I am analysing netCDF data using IDL or a simple Fortran program I
wrote myself, or working on a laptop which doesn't have a cache and isn't
connected to the internet, I need a netCDF file which contains the attributes
that describe the coordinate ref system.

However, I accept the argument that URNs may give useful access to extra
descriptive metadata which is not essential to analysis or plotting.

Best wishes

Jonathan
Received on Thu Oct 04 2007 - 04:50:14 BST