This is a matter of philosophy and practice, I imagine.
One of the reasons I recommend semantic URIs is that the URI
represents a clue, if by some chance the web is not available or the
information is not served at that URI any more. I think some semantic
description should be included in the netCDF file -- on this I agree
with Ed.
But I agree with Roy on the bigger picture -- we are building a system
that is intended to be long-lived, and in fact much of the web is
moving to systems that assume URLs are long-lived. There are a lot of
problems that will result if vocabulary servers like Roy's go away.
And the fact is, you just can't store all the metadata in the file --
there is an expanding web of information associated with every
observation, and much of that metadata can evolve as our knowledge
evolves about the relevant instruments, best calibration values,
algorithms, and so on. So we need to allow the ability to reference
additional metadata, not just bring it all into every netCDF file (and
then update all the files when additional metadata is obtained).
John
On Oct 27, 2009, at 0837, Ed Hartnett wrote:
> <martin.juckes at stfc.ac.uk> writes:
>
>> Hello Roy,
>>
>> That is an interesting idea. There are definitions of these areas
>> on a WWF site, of the form
>> http://www.worldwildlife.org/wildworld/profiles/terrestrial/nt/nt0908_full.html
>> where "nt0908" is a tag associated with a particular flag value.
>>
>> My first thought was:
>> URI=" http://www.worldwildlife.org/wildworld/profiles/terrestrial/
>> [flag_meaning]_full.html" with "nt/nt0908" in the flag_meanings
>> attribute, but it appears that the cf-checker does not like having
>> a "/" in flag_meanings. This means we need to either have a more
>> complex syntax, or set up our definition URLs along the lines you
>> suggest.
>>
>> A more complex syntax might be:
>> URI=" http://www.worldwildlife.org/wildworld/profiles/terrestrial/
>> [flag_meaning][0:2]/[flag_meaning]_full.html"
>>
>> Since the above returns html rather than skos/xml it should perhaps
>> be complemented by a uri_content_type attribute, with value 'html/
>> text'.
>>
>> I'll think about the option of setting up something like your vocab
>> server -- or possibly putting some more definitions in the ndg
>> vocab server?
>>
>> Regards,
>> Martin
>
> Howdy all!
>
> I really don't think you should store metadata for a file on the
> web. Just put it in the file.
>
> Most scientific data are around a lot longer than most URLs. What
> happens in 15 years when someone wants to understand this data, and
> the
> URL has vanished? It the metadata were stored in the file, then it
> would
> be there for that programmer of the future. (And that programmer may
> well be you! <insert scary Halloween laughter here>)
>
> If the metadata seem too large to put in the file, consider storing
> them
> in variables, and then use netCDF-4 with compression turned on for
> those
> variables.
>
> Thanks,
>
> Ed
>
>
>
> --
> Ed Hartnett -- ed at unidata.ucar.edu
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
--------------
I have my new work email address: jgraybeal at ucsd.edu
--------------
John Graybeal <mailto:jgraybeal at ucsd.edu>
phone: 858-534-2162
Development Manager
Ocean Observatories Initiative Cyberinfrastructure Project:
http://ci.oceanobservatories.org
Marine Metadata Interoperability Project:
http://marinemetadata.org
Received on Tue Oct 27 2009 - 10:21:13 GMT
