[CF-metadata] Getting back to ensembles

[Francisco Doblas-Reyes]

Hi Jamie,

Thanks for the comments.

I remember the discussion on the need to flagging provisional files. 
However, we haven't advertised this server for official dissemination, 
so I didn't think the flagging was necessary at this stage. We use the 
OPeNDAP server only for testing (for instance, to check which software 
is able to access, read and display 5-dim fields from an aggregated 
server). That is also the reason why I decided to keep the 
standard_names of the variables not yet included in the list, so as to 
have examples with all the required elements.

Paco



Kettleborough, Jamie wrote:
> Hello Paco,
> 
> thanks for giving us sight of these - do you care that they are not,
> strictly speaking, CF compliant as they use standard_names (like source,
> etc) that are not yet in the standard_name list?  Should the
> standard_names just be dropped for now and rely on the long_names?
> 
> Though I think this is all tied up with process (apologies for mixing
> threads).  If we are going to have 'test files' and test implementations
> don't we need some way of flagging the bits in the files that are
> provisional or being tested?
> 
> Jamie
> 
-- 
________________________________________

Francisco J. Doblas-Reyes
European Centre for Medium-Range
Weather Forecasting (ECMWF)
Shinfield Park, RG2 9AX
Reading, UK

Tel: +44 (0)118 9499 655
Fax: +44 (0)118 9869 450
f.doblas-reyes at ecmwf.int
_______________________________________