At 3:19 PM +0100 10/3/07, Jonathan Gregory wrote:
>To some extent I am being devil's advocate here, but I would like to know what extra information you get about the CRS if you look up the URN than we are anyway proposing to store in netCDF attributes.
Ah, but there are so many bits of extra information that may be useful or interesting in various ways. (Admittedly not directly required for immediate calculations of the file's data, but still useful in other ways.) A few examples:
- It is much easier to compare a unique ID than to compare all the attributes necessary to describe the entity. (Corollary: It is much easier to spot differences when the CRS is different.) This is true for both humans and computers, and for the programmers that program the computers.
- Reference information for the new user of a particular CRS is much more readily available via the URN.
- Contextual information is available via the URN. (When did this CRS get standardized, if at all? By whom? Who uses this CRS?)
- Labelling the results (e.g., with ancillary metadata about the CRS) is possible with the additional metadata.
Saying the attributes have all the information needed to process the file presumes a fairly narrow definition of the term 'process', one that doesn't allow for many future applications.
I doubt we should ever get rid of the attributes, but I don't think one could ever think the file is *fully* self-describing just because of the attributes currently defined.
John
--
----------
John Graybeal <mailto:graybeal at mbari.org> -- 831-775-1956
Monterey Bay Aquarium Research Institute
Marine Metadata Initiative: http://marinemetadata.org || Shore Side Data System: http://www.mbari.org/ssds
Received on Wed Oct 03 2007 - 09:34:59 BST
