[CF-metadata] 2. Re: [cf-satellite] Sharing quality flags among multiple variables (Jonathan Gregory)

From: Jonathan Gregory <j.m.gregory>
Date: Mon, 21 Nov 2011 17:59:46 +0000

Dear Martin

I am sure that most things could be done differently and many of them could
be done better, but we have to recognise that few people have much time to
spend on defining the CF standard, because it's a community effort with very
little devoted resource. We can certainly consider ways to reform it, but it
is slow, as you said. The annual GO-ESSP meeting has often/usually had a "CF
day", which is an opportunity for a meeting. In practice, I think there is
actually far too much to discuss. Many of our discussions require a lot of
thought, and I suspect they are hard to settle at a one-off meeting. I tend
to favour teleconferences for specific points.

Also, as your elaboration suggests, one can imagine lots of ways in which
things can become immensely complicated. For me, that is the main reason for
the principle that we only introduce new conventions when we have a definite
need for them, rather than trying to foresee and allow for all possible
generalities. Being concrete limits the complexity of the convention.

> Therefore, I would argue that the "synchronizing" of obs is a special thing and should be treated separately from the statistical treatment of variables. Hence, the default in the sea water temperature and salinity case would be that each variable has its own "count" (via cell_methods?). One could then define a way to create this link via some sort of cross-referencing.

I think I agree with that, and isn't it the situation we have? The statistical
treatment of variables is done by cell_methods; the metadata for obs is done
by standard_name modifiers, which are separate for each variable, though we
are also discussing on this email list a way to link them together if needed.

Best wishes

Jonathan
Received on Mon Nov 21 2011 - 10:59:46 GMT