Hello everyone,
During a recent discussion, it transpired that there are conflicting
rules regarding the "Conventions" global attribute:
* The CF conventions state that its use is _recommended_
(
http://cf-pcmdi.llnl.gov/documents/cf-conventions/1.6/cf-conventions.html#identification-of-conventions)
* On the other hand, the conformance document states that its use is
_mandatory_ (2.6.1 in
http://cf-pcmdi.llnl.gov/conformance/requirements-and-recommendations/1.6/)
We felt that the conformance document ought to be correct, i.e. the
"Conventions" global attribute must be set, and the conventions
document is wrong.
An example of when not setting "Conventions" could cause real
confusion is with the use of attributes that were non standard in
older versions of CF but are standardised in later versions. For
example, the following data variable:
byte sensor_status_qc(time, depth, lat, lon) ;
sensor_status_qc:long_name = "Sensor Status" ;
sensor_status_qc:valid_range = 1b, 63b ;
sensor_status_qc:flag_masks = "QWERTY";
is complaint at CF-1.2, but not at CF-1.6. This is because the
"flag_masks" attribute used not to have any special meaning.
I would like raise a defect trac ticket to fix this (either way), but
would welcome any comments here, first.
All the best,
David
--
David Hassell
National Centre for Atmospheric Science (NCAS)
Department of Meteorology, University of Reading,
Earley Gate, PO Box 243,
Reading RG6 6BB, U.K.
Tel : +44 118 3785613
E-mail: d.c.hassell at reading.ac.uk
Received on Thu Oct 03 2013 - 03:16:44 BST
