[CF-metadata] Rotated-pole grids

[Bentley, Philip]

Hi Jon,

As you probably recall, my CF Trac tickets #9 and #18 stimulated a good
deal of online debate regarding which CRS attributes to incorporate into
the CF conventions. In the end a much reduced set of attributes was
added, primarily because we couldn't reach agreement on some of the more
complex CRS properties. Although this was disappointing from a personal
perspective, at least we got some new features added!

As for improving CF/netCDF-to-GIS interoperability, my programmer nature
says to exploit standardised CRS identifiers (such as those maintained
by OGP/EPSG). However, there was such resistance to using opaque codes
like these within CF (the old self-describing argument) that I think
this will continue to be a hard sell.

Thus, IMHO, the CRS WKT (well-known text) route appears to offer the
most promise, even though this too has limitations. I think that's what
I concluded towards the end of one of the aforementioned tickets.

Cheers,
Phil

> -----Original Message-----
> From: cf-metadata-bounces at cgd.ucar.edu 
> [mailto:cf-metadata-bounces at cgd.ucar.edu] On Behalf Of Jon Blower
> Sent: Wednesday, June 17, 2009 7:50 PM
> To: John Caron
> Cc: Adit Santokhee; CF metadata
> Subject: Re: [CF-metadata] Rotated-pole grids
> 
> Thanks all - I can certainly see the convenience value in 
> including the lat/lon arrays, but it is frustrating that some 
> tools will reject a dataset because of the absence of 
> mandatory fields that they don't need!  Still, you have 
> provided many valid arguments for retaining the arrays and 
> it's good to have confirmation that this information is 
> indeed redundant.
> 
> Just a word about GIS clients, picking up on one of Steve's 
> comments - many GIS clients would prefer to have the 
> coordinate system properly specified (as a CRS code or 
> Well-Known Text) rather than by listing the lat-lons 
> exhaustively.  (Another conversation could be started to 
> discuss the value of including these as optional attributes, 
> since most of the CF projections are commonly used in GIS.)
> 
> Cheers, Jon
>