Re: Thoughts on GALEON Phase 2

  • Subject: Re: Thoughts on GALEON Phase 2
  • From: John Caron <caron@xxxxxxxxxxxxxxxx>
  • Date: Wed, 01 Mar 2006 09:21:38 -0700
Woolf, A (Andrew) wrote:

Here, the 'xlink:role' attribute is used to indicate an assumed
well-known procedure for dereferencing netCDF Grid axes to a GML
domainSet. This - as well as having WCS provide netCDF as an output
format - is only possible if 'netCDF' has explicit well-defined
semantics that allow it to be interpreted. But it doesn't! Hence our
recommendation on CF-compliant netCDF as an additional WCS format.

I agree and would add that we need 1) to create a document explicitly 
specifying the mapping of CF/Netcdf to WCS (if its trivial, it will be short!); 
2) we need to extend CF in various ways needed for high resolution geolocation, 
for example specifying Datum; and 3) we need to modify CF in a small number of 
ways, eg allowing a valid CF file to not contain a lat/lon array.

I recommend we create a subgroup who are interested in this, and then actually do it!