Hi Karl, Glenn, Nan and CF-Metadata list, I feel that recommendation could be expanded upon (added bit Italics below) to say:
3 Description of the Data Recommendations: a.. All variables should use either the long_name or the standard_name attributes to describe their contents. Exceptions are boundary and climatology variables. Use of the standard_name is strongly recommended for inter-operability and metadata discovery purposes. However if the standard_name does not suit the data then use a long_name. Andrew Walsh Data Facilitator AODN ----- Original Message ----- From: Karl Taylor To: [email protected] Sent: Saturday, May 28, 2011 4:54 AM Subject: Re: [CF-metadata] CF Convention - standard_names Hi Glenn, My reading of the conformance document (as well as the conventions document) is that standard_name is optional (but either standard_name or long_name is recommended). Here's the relevant section from the conformance document: 3 Description of the Data Recommendations: a.. All variables should use either the long_name or the standard_name attributes to describe their contents. Exceptions are boundary and climatology variables. I agree with Nan, that it is essential that variables without standard names be allowed. regards, Karl On 5/27/11 7:34 AM, Nan Galbraith wrote: Hi Glenn - Given that this would greatly undermine the purpose of the CF Convention, On the contrary, it's very important that variables without standard names be allowed in CF files. In OceanSITES, we use this "feature" of the convention to allow users to include instrument-specific and provenance-related variables. These are typically variables that do not need to be discoverable, and often don't need to be understood outside the community in which the data is generated. We require standard names for "scientific observables" but encourage people to include other variables that help describe the data collection, quality, and processing of those observations. To require standard names for all variables would make CF untenable, and would drown this list in requests for new names. Or worse, it would encourage data providers to omit key metadata that's best expressed as a variable in NetCDF. Regards - Nan Galbraith On 5/27/11 7:25 AM, Comiskey, Glenn wrote: Hi, I note a discrepency between the CF document and the CF conformance requirements and recommendations document, for all versions v1.0 thru v1.5, in that the CF conformance requirements and recommendations document states that the standard_name attribute is a requirement, while the CF document details standard_name as "an optional mechanism" for describing the data being represented. The CF conformance requirements and recommendations document clearly states that in the event of "any discrepancies between the two, the conventions document is the ultimate authority." This would seem to imply that standard_name is in fact not a requirement for CF conformance. Given that this would greatly undermine the purpose of the CF Convention, i.e. "provide a definitive description of what the data...represents", I am assuming that the authors of the CF Convention intend for standard_name to be mandatory for CF conformance. However, I am hoping if someone is able to definitively state whether the standard_name attribute is a requirement or is optional, i.e. recommended, but not obliqatory. I dare say this matter has been raised before, but having only just started working with metorological/oceonographic data sets I'm not aware of what the accepted answer to my query is; and I haven't been able to find anything in the archive relating to this matter. Kind regards, Glenn Comiskey Data System Administrator _______________________________________________ CF-metadata mailing list [email protected] http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata ------------------------------------------------------------------------------ _______________________________________________ CF-metadata mailing list [email protected] http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
_______________________________________________ CF-metadata mailing list [email protected] http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
