Hi Bryan,
Hi everybody,
I had not anticipated such a discussion on this topic. In my original post I
had left the direction of the final proposal to the CF community. At that
time I mentioned three possible approaches: 1a) dummy quantity that refers
to its components, 1b) reference to other components via attributes, and 2)
standard prefixes. The discussion has made clear that the option 2 is
definitely not the way to go. However, I am not sure whether the preferred
approach is to go for:
* as 1a): an additional dummy variable of some type "vector" (or tensor)
that refers to its components.
* or 1b): an attribute associated with each component that refers to the
other components.
Are there any existing extensions or proposals that I should check for the
proposal? Anything related to virtual or multiple related files, or order of
components (Z Y X)?
Best regards,
Bert
----- Original Message -----
From: "Bryan Lawrence" <b.n.lawrence at rl.ac.uk>
To: <cf-metadata at cgd.ucar.edu>
Cc: "John Caron" <caron at unidata.ucar.edu>; "Jonathan Gregory"
<j.m.gregory at reading.ac.uk>
Sent: Tuesday, March 21, 2006 8:57 AM
Subject: Re: [CF-metadata] CF convention for vector quantities
>
> Hi Jonathan et al.
>
> Jonathan said:
> > > I remain unconvinced about introducing this kind of information into
the
> > > file, as it cannot be depended upon if it is optional, so it doesn't
help
> > > simplify programs, and it is redundant because it can be deduced on
the
> > > fly (I think). However I know I have a general preference for clever
> > > programs and stupid (though complete, correct and clear) files.
>
> John said:
> > I appreciate your advocacy of that POV throughout the years, its very
> > important that we maintain CF files to be as simple amd clear as
possible.
> > I also observe the tension between data providers and consumers, and it
> > seems likely we will always be looking for the right balance between
> > program writers and file writers.
>
> and despite the disagreements that Jonathan and I have , I'd like to echo
> this. I think it's healthy to have a tension, so we can seek the right
> balance, and that depends on consistent well argued positions from all
> sides :-)
>
> Which brings us back to "can we come up with a simple yet useful
proposition"
> that doesn't overload the standard names (in any way, I totally agree that
> this concept should not touch them) ...
>
> ... if no one else does so before then, I've allocated some time in a
couple
> of weeks to come up with a proposal and an example file, and a decent
> justification ... (Bert, I hope you get to it before me :-)
>
> Thanks
> Bryan
>
> --
> Bryan Lawrence
> Director of Environmental Archival and Associated Research
> Head of the NCAS/British Atmospheric Data Centre
> CCLRC, Rutherford Appleton Laboratory
> Phone +44 1235 445012; Fax ... 5848; Web: home.badc.rl.ac.uk/lawrence
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://www.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
Received on Tue Mar 21 2006 - 01:41:34 GMT
