Hi Brian,
I guess part of the exercise is to reach an agreed upon terminology!
I don't think of station data and trajectories as having boundary
information, instead being defined at specific lat/lon/time points with no
associated cells.
So in typical usage would you expect station data and trajectories to have
associated bounds?
Bob
-----Original Message-----
From: Brian Eaton [mailto:eaton at mscan.ucar.edu]
Sent: Tuesday, June 18, 2002 4:06 PM
To: Bob Drach
Cc: cf-metadata at cgd.ucar.edu
Subject: Re: [CF-metadata] CF question
Hi Bob,
Seems that determining when a variable is "gridded" depends on the
attributes of the GriddedVariable class. I consider station data and
trajectories to be fine examples of gridded data, so when you say that they
don't qualify as GriddedVariables, that indicates a GriddedVariable must
have attributes that station data and trajectories don't have. What are
they? Can you use them to distinquish gridded and non-gridded variables?
Brian
On Tue, 18 Jun 2002, Bob Drach wrote:
> All,
>
> A question on CF arose in the context of the Earth System Grid project.
> We've started to design a database for CF/NetCDF metadata, and are at
> the point of developing an object model for the database. One of the
> classes is GriddedVariable, that is, a variable defined on a spatial or
> temporal grid. Station data and trajectories would not be considered as
> GriddedVariables.
>
> The question centers around how an application, say a metadata
> harvesting utility, would recognize that a variable is gridded. One way
> would be to check if the associated coordinate variable(s) have boundary
> variables. But the concern was raised that CF data producers might opt
> not to store boundary variables, either because it's unimportant for
> their application or it's too much trouble. In such cases it would not
> be possible to infer that the data was in fact gridded.
>
> So the question is: should an application assume that a variable is
> non-gridded if it is missing explicit boundary data? Or would it be
> better to require that nongridded data be tagged in some fashion, with
> the expectation that most CF data will be gridded, and infer that a
> variable is gridded unless otherwise indicated?
>
> Bob
>
>
>
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://www.cgd.ucar.edu/mailman/listinfo/cf-metadata
Received on Tue Jun 18 2002 - 19:12:59 BST
