Tom and other cf-satelliters,
> Each "band" may have its own "unit" or even scale and offset.
> Presently, there are no provisions for a "dimensioned attribute"; I
> supposed these attributes could be specified as variables, but I worry
> about applications trying to connect them (or the library, in the case
> of scaling and NetCDF-Java).
I think variables would be adequate for this. What connects them to the
bands is the shared band dimension.
> I'm hoping the NetCDF developers have thought about this...and can
> steer us a bit into what makes sense.....
Attributes can also have multiple values, but not using one of the
shared dimensions that variables use to specify shapes and shared grids.
In netCDF-3, multiple attribute values are used to implement strings,
since the NC_CHAR primitive type is only a single character.
In the enhanced netCDF-4 data model, strings are a primitive type, so an
attribute may have as values multiple strings of different lengths.
This could also be provided in the netCDF-3 data model with a convention
for how to delimit multiple string values in a single netCDF-3 text
attribute. Since netCDF-3 attributes have a length, the null character
could be used for this purpose, but CF conventions tend to just use a
space character to delimit multiple strings such as names in a single
text attribute. For example, see the CF attributes for coordinates,
bounds, cell_measures, cell_methods, and formula_terms.
--Russ
