[CF-metadata] quantities requiring parameters to define them

From: Jonathan Gregory <j.m.gregory>
Date: Tue, 26 Oct 2004 23:25:45 +0100

Dear Brian

Initially I thought that formula_terms *was* doing the same thing as what we
need here, but your objection pointed out that it wasn't. Given that neither
existing mechanism is exactly what we want, let's have a new one, as you say.
I agree that having an attribute meaning different things according to context
is confusing.

What about an attributes of "parameters"? That would be a parallel to the
attribute of "coordinates", especially thinking of scalar coordinates. Both
parameters and coordinates are independent information on which the quantity in
the data variable depends. The difference is that coordinates have a meaning
independent of the quantity in the variable, being either spatiotemporal
information or measurable quantities, whereas parameters are supplied solely to
define the quantity. Examples are reference pressure, difference in temperature
or density from the surface for the ocean mixed layer, and perhaps the identity
of the reference ellipsoid (from another thread).

The parameters attribute would have the "keyword: value" format of formula_terms
with the standard_name definitions specifying what keywords and values are
permitted.

Cheers

Jonathan
Received on Tue Oct 26 2004 - 16:25:45 BST