[thredds] NCML in aggregations

[David Neufeld <David.Neufeld@xxxxxxxx>]

Hi All,

I'm trying to better understand how NCML attributes are populated in an 
aggregated dataset.  For example, is the first netdcdf file used as the 
'template' from which global attributes are populated (alternatively 
perhaps some random sample is used)?

This is in the context of generating NCML from an aggregated THREDDs 
dataset.

Thanks, Dave