[CF-metadata] physical vs dimensional units

From: John Caron <caron>
Date: Wed, 13 Apr 2011 20:26:00 -0000

On 4/13/2011 9:24 AM, Jonathan Gregory wrote:
>> > I do sympathize with people who prefer to read "nmol mol-1" on a plot rather than "1" (or "1.e-9" in this case).
> Since nmol mol-1 is equal to 1e-9, there is no problem with using it as units
> in a CF file. It is udunits-compliant. It could even be stated as the
> canonical unit in the standard name table.
>

the point im trying to make is that it would be better to understand
that "mol mol-1" (canonical udunit = 1) is not the same as "m3 m-3"
(canonical udunit = 1).

if you prefer to keep the word "unit" to mean "dimensional unit", then
we could use another term, say "physical quantity" or something.

BTW, we might want to clarify what "udunits-compliant" means. the java
udunit package appears to accept the string "garbage garbage-1". Im not
sure about the udunit C library.
Received on Wed Apr 13 2011 - 14:26:00 BST