[CF-metadata] variables in separate files

From: Jonathan Gregory <jonathan.gregory>
Date: Sun, 6 Apr 2003 10:01:58 +0100

Dear All

Brian has remarked:

"I don't think CF needs to address all possible ways that someone might
choose to distribute data between files. For example, we don't provide
conventions to store coordinates in a separate file from the data
variables."

I agree that we can't foresee how the data might be separated. However, I
do think we ought to acknowledge that it will happen. The need to put
coordinates in a separate file has come up many times, and people will
certainly do it. Also, I believe that error variables are likely to end up
sometimes in separate files.

That means we have to design the standard with this possibility in mind. In
general I think it means variables should be independently self-described.
I also think we should say *something* about it in the standard, simply in
order to acknowledge that it is allowed/necessary, if nothing else. Although
the standard is for netCDF files, we are all aware that the design of the
metadata is of more general applicability, including situations where the data
is not stored in named variables in discrete files. I haven't got a proposal
for what to do say, but I'd like to raise this issue as a separate thread to
see what people think about it.

Best wishes

Jonathan
Received on Sun Apr 06 2003 - 03:01:58 BST