Hi Jonathan
Not sure what's behind your question. Do you have a problem with
versioning per se?
I think one wants to think carefully about names not being withdrawn.
Quite clearly one can never withdraw the fact that someone used a name
with an associated definition, but one can and should envisage that in
time the use of a specific name should be deprecated and then
withdrawn/replaced for *current* use. This is one specific use case
which we do need to support with versioning.
For example, imagine that the definition of some specific quantity was
found to be wrong? What would we do then? This is almost a dead cert to
happen as we expand from things which are physical measurements on to
chemical measurements and classifications (with no disrespect to anyone,
somethings are inherently defined and some are not).
Further, one might in time add specific further detail to a quantity,
allowing more precise standard names. For a rather contrived example,
perhaps we might start with chemical definitions which are for a family
of molecules. We might not at the time include standard names for all
the constituents. Someone wanting to use a standard name might use the
family name (which is all that is available). That should be versioned,
so that in the future when perhaps we have gotten around to standard
names for all the constituents, we can read this file, and know that the
originator did not have the facility to be more precise, and not
therefore overinterpret the data.
Regards,
Bryan
On Thu, 2006-09-21 at 21:28 +0100, Jonathan Gregory wrote:
> Dear Roy
>
> Versioning of standard names: Once introduced, they are never withdrawn, though
> they might be made into aliases if a mistake is discovered or a change of mind
> takes place. What version info do you need?
>
> Cheers
>
> Jonathan
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://www.cgd.ucar.edu/mailman/listinfo/cf-metadata
Received on Fri Sep 22 2006 - 01:02:36 BST
