[CF-metadata] what standard names are for

From: Jonathan Gregory <j.m.gregory>
Date: Sat, 12 Apr 2008 11:04:01 +0100

Dear Craig

Yes, I've already said that I agree that your list

> surface_temperature (already defined)
> sea_water_temperature (for temperatures at depth)
> sea_surface_skin_temperature
> sea_surface_subskin_temperature
> sea_surface_foundation_temperature

is in the end the best solution we can agree on, although it comes full circle,
as you said.

I am sorry that I subsequently confused this or it got lost by reusing this
example in the wider debate on what standard names are for, because it's a
good example. The discussion is important but does not solve any immediate
problem, so we should go with what you have proposed.

I am going to make further comments now, but that doesn't mean I am arguing
against the above!

You are quite right that to serve the community better CF should make its
decisions faster. This is a long-standing problem, which partly comes from the
fact that despite its growing importance, and the fact that it is used by many
projects (like yours) to which large resources are committed, CF itself still
has very little dedicated staff effort. It is a common resource which benefits
many institutions but is paid for by hardly anyone. (That's an even wider issue
than standard names in general, so doesn't really belong in this thread!)

I tried to explain in an earlier posting why I think standard names are partly
definitions, and why that means they are not just "names". It is because the
community the data serves is often broader than the one which produces the
data. What is understood in a smaller community is often not so clear to a
larger community. That doesn't imply any disrespect to the authority or the
expertise of the smaller community. I understand that people may feel indignant
if they think someone else is claiming to describe their data better than they
can themselves, but that is not the intention. When I have questioned standard
name proposals, the question is often, "What does that mean?", and then I have
often made suggestions that the standard name adopted should have elements of
the *answer* to the question, so that other users will not have to ask. They
can just look at the metadata and know what the data is. As a data analyst, I
think this is very important. I don't want to have to spend a lot of time
hunting down definitions and documentation to decide which of the quantities in
a dataset is the one I want to use. I hope that seems reasonable to you.

Thank you for persisting with CF.

Best wishes

Jonathan
Received on Sat Apr 12 2008 - 04:04:01 BST