Hello again,
In my posting of 28 Oct, I mentioned that I have written a tool for
generating an XML-based markup of CF-metadata Standard Names according
to a grammar based on the published Guidelines for constructing such
names. This tool necessarily parses the names along the way, to ensure
that they can be correctly converted to the XML format.
I have now processed all 1103 entries in the Standard Name table, and
have found that just 25 parse correctly according to the grammar. An
example of a Standard Name that does parse correctly is
downward_eastward_stress_at_sea_ice_base
and one that doesn't:
air_temperature.
This low success rate is not particularly surprising, since:
- The Guidelines are just that, and are not (I presume) meant to be the
definitive grammar specifying how names are constructed (although the
Qualifications and Transformations sections are pretty close to being
grammar rules).
- Two of the tables included in the Guidelines (Special Phrases and
Chemical Species) are not (as far as I can see) referenced in the
Guidelines's grammar-like rules, so have been left out of my initial
crude attempt to come up with a grammar.
- There will always be a need for special terms which do not fit into
any standard grammar.
- There could be errors of understanding or implementation in my draft
version of the grammar (though I have looked at some of the failed
Standard Names, and can see why they failed).
I'd like to ask whether there is any interest (or past or current work)
within the CF-metadata initiative to develop a formal grammar for at
least some proportion of the Standard Names. If so, I'd be happy to
share the work I've done so far. This consists of two versons of the
grammar - one in XSugar (as per my email of the 28th Oct), and one in
Prolog, which uses a rather more readable grammar notation. Please
feel free to contact me off-list.
Cheers,
Robert
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
_______________________________________________
CF-metadata mailing list
[email protected]
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
