[CF-metadata] GIS versus CF conventions

From: Jonathan Gregory <j.m.gregory>
Date: Tue, 30 Aug 2005 08:15:31 +0100

Dear John

If I was someone whose coordinate metadata took 15/16 of the file, I would
certainly want to leave it out! Therefore I think it's inevitable that we
should do something like this:

> I would propose some standardized "CF variants", or profiles, to be worked
> out by subgroups that need variations from the standard, rather than
> modifying the base standard. The groups should have a clearly stated target
> or purpose, perhaps as narrow as "WCS" or as broad as "GIS".

one such variant being the omission of auxiliary coordinate variables if
appropriate map projections or grid descriptions have been provided so that
the coordinate variables can be calculated from the indices. However it would
be good to:

* Maximise interoperability by providing the coordinate variable information
in some other file. Could that be done in the case you're talking about?

* Emphasise that this route should only be taken when the coordinate metadata
really takes an intolerable amount of space. We shouldn't make auxiliary
coordinates generally optional. I'm not sure how we can do this, but as you
say there has to be some clearly stated purpose. "GIS" sounds a bit vague to
me even. Could we relate these provisions to the needs of specific projects
or software?

Best wishes

Jonathan
Received on Tue Aug 30 2005 - 01:15:31 BST