Hi John,
We label Units of Measure with a URI (e.g. SDN:P061::ULAA represents metres), which represents a term in the P061 vocabulary (
http://vocab.ndg.nerc.ac.uk/P061/list/current). The vocab (born in the late 70s) is a work of pragmatism that contains a mixture of 'dimensional units' plus 'physical units' carrying semantics like 'mg/g dry sediment'. Work in progress is to migrate semantics into our equivalent to Standard Names deprecating P061 terms as we go with the medium term objective of getting as close to udunits as possible. However, I feel complete harmonisation will never be possible.
All our current software does with the P061 URIs is to use them to create text labels for plots, ASCII listings and so on. The resulting consistent labelling is the primary objective of P061. Of course, my ambition is to develop UoM interconversions, probably built on udunits, but that isn't even at the vapourware stage yet. So, the answer to your question is that we use 'physical units' as labels, but don't try and do any processing based on them.
I'm starting to wonder whether the 'labelling' and 'interconversion' use cases have become a little confused in the discussions of the past few weeks.
Cheers, Roy.
-----Original Message-----
From: cf-metadata-bounces at cgd.ucar.edu [mailto:cf-metadata-bounces at cgd.ucar.edu] On Behalf Of John Caron
Sent: 31 March 2011 14:00
To: cf-metadata at cgd.ucar.edu
Subject: [CF-metadata] physical vs dimensional units
udunits is a "dimensional units" library, for manipulating powers of the
fundamental dimensions (length, mass, time, charge, temperature). this
is necessary but not complete for capturing the meaning of the "physical
units" of our data.
We also need units like kg/kg, for which the udunit canonical string is
the empty string. But even more difficult is "atoms CO2 / atoms air" or
"grams CO2 / grams air", or "count of phytoplankton".
the simple thing is to just introduce another attribute "unit labels"
(or something) for this, to be displayed to the user. but such an
"opaque string" limits the amount of automatic inferencing that can be
done. better would be a grammar from which both the dimensional units
and the "substance we are talking about" can be understood.
also, all of our space/time units arent dimensional units, they are all
referenced to a datum. we include the datum in the udunit string for
time, but not for vertical or horizontal coordinates. thats not a
particular problem, but it does point out that these units are not the
same as dimensional units. we need to include the datum in the "physical
units" representation, which could be one string or several strings .
what are other solutions that are being used for "physical units"? Im
wondering how Roy Lowry's software deals with this? Or the MMI or SWEET
ontologies , etc? Is there something nice and compact like udunits
strings, but with more semantics, without getting into the complexity of
RDF triples?
_______________________________________________
CF-metadata mailing list
CF-metadata at cgd.ucar.edu
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
--
This message (and any attachments) is for the recipient only NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.
Received on Thu Mar 31 2011 - 07:39:17 BST
