CF Standard Names Status and Development Alison Pamment Alison Pamment [email protected] [email protected]
Jan 17, 2016
Overview Current status of CF standard
name table Process for agreeing standard
names Development needs
Current Status Currently 1039 names in table Represents approx 30% increase
in size of table over 12 months A further 40 proposed names are
agreed and ~30 are currently under discussion
Standard Name Process (1) All proposals for new standard names
must be made via the CF-Metadata mailing list
There are published guidelines for the construction of standard names
Proposals should include the unit and a definition
Standard Name Process (2) Discussion of proposals takes place
on the mailing list and all may participate
The aim is to achieve consensus The standard names manager acts as
moderator and keeps the discussion moving forwards
Standard Name Process (3) Names for which consensus has
been achieved are flagged as ‘accepted’ by the moderator Accepted names will go into the table at the
next bimonthly update Accepted names can be used immediately –
don’t have to wait for table update to take place
Standard Name Process (4) Where consensus cannot be reached
by discussion the moderator will ask the standard names committee to vote on the proposed names
The committee’s decision will be final All versions of the standard name
table and guidelines will be kept online
Development Needs – the problem Many standard name proposals are pretty
straightforward…BUT A minority of name proposals are
controversial and raise more questions than they answer
FOR EXAMPLE: For new name proposals how much
weight should be given to current usage in communities beyond CF?
Development Needs – the problem Should we use human readable
strings or opaque URIs that point to a definition?
Should we put the definition in the file?
Are standard names too long to be useful?
We need a standard names ‘philosophy’
Standard Name Philosophy Are the names aimed primarily at the
proposing community or to be easily understood by the whole CF community?
Human readable names certainly shouldn’t be misleading
CF standard names can never be all things to all people
Development Needs
How do we express the relationship between CF standard names and external vocabularies?
Currently we have GRIB and AMIP equivalences but they are out of date
NDG vocabulary server Tickets 24 + 27 + 29 + … :
‘common_concept’
Common_concept The proposal is to allow a ‘bundling’ of CF
metadata to include standard_name + other attributes, e.g, 2m temperature
The ‘bundle’ could consist of a standard_name alone
The bundle would have a scoped name and a URI that points to its definition
Need a way of defining the constraints on a common_concept definition (#29)
The URI could point to metadata not defined in CF (#27)
Common_concept
Registration of a common_concept would be quick (automated)
Another user could register a different scoped name for an existing common_concept but the URI would be identical
Common_concept would facilitate: The creation of short names Translation between standard_names and
other vocabs
Other Development Needs Instrument information – should it go
in the standard name? Chemistry names (IUPAC?) Ticket 17: use cell_methods instead
of ‘where’ names Ticket 33: ‘count over days’ for use
with standard names for climate statistics
Summary Agreed rules for standard names
proposals will be published on CF website
We need to develop a clear standard names ‘philosophy’
The common_concept idea has received unanimous support in principle – there is much work to do on the practical implementation