⇐ ⇒

[CF-metadata] per-variable metadata?

From: Lowry, Roy K. <rkl>
Date: Sat, 6 Aug 2011 08:21:56 +0100

Hi Nan,

I admit that my thinking is driven by coming from a community (SeaDataNet/Geo-Seas) where 99% of data are currently in formats other than netCDF (hopefully something that will change in SeaDataNet II) and therefore need a usage metadata solution that will initially at least operate across all formats in use. This is why I'm looking at O&M/SensorML. I have real concerns that a scenario will develop where SDN II partners build XML usage metadata stock for their ODV (the ASCII format holding most SeaDataNet data) data holdings and then resist migration to CF netCDF if this XML stock cannot be utilised.

Whilst admittedly this isn't 100% relevant to a discussion on CF, I would argue that if a universal standard requiring XML encoding is developed CF should buy into that rather going its own way. Note that a solution where a specified binding to external resources is defined that can sit side-by-side with internal encodings for a usage metadata subset would make me happy and shut me up (on this issue at least).

Cheers, Roy.
________________________________
From: Nan Galbraith [ngalbraith at whoi.edu]
Sent: 05 August 2011 20:58
To: Lowry, Roy K.
Cc: John Graybeal; cf-metadata at cgd.ucar.edu
Subject: Re: [CF-metadata] per-variable metadata?

I don't think the metadata structure needs to be terribly
complex, in most cases. Changes to instruments (either a
new instrument, a change to the sample scheme, or some
kind of change to calibrations, etc) partway through could
be a problem, but that's true in xml too.

John G is right, we often lose access to the network at sea;
but even without that consideration, NetCDF is supposed to
be self-documenting, and we should be able to come up with
some strategies for including fairly complex metadata in the
files.

I like the ancillary variable route because it seems straightforward
and flexible.

 * Data variables can share one or more ancillary variables, very useful for something like an ADCP record.
 * Ancillary variables can have data types.
 * Ancillary variables can have dimensions that make them self-explanatory.

Cheers - Nan


On Aug/05/2011 2:08 PM, Lowry, Roy K. wrote:

Hi John,

We seem to have a difference of perception in the fragility of Web access to the average scientist. Certainly, I was assuming ships had virtually the same level of internet access as I have at home (if not in the lab) based on my experience of the NERC ships. However, if lack of internet access becomes an issue then we need to look at URI resolution to physical XML files sitting with the NetCDF.

Cheers, Roy.

________________________________________
From: John Graybeal [jbgraybeal at mindspring.com<mailto:jbgraybeal at mindspring.com>]
Sent: 05 August 2011 17:57
To: Lowry, Roy K.
Cc: ngalbraith at whoi.edu<mailto:ngalbraith at whoi.edu>; cf-metadata at cgd.ucar.edu<mailto:cf-metadata at cgd.ucar.edu>
Subject: Re: [CF-metadata] per-variable metadata?

On Aug 5, 2011, at 09:27, Lowry, Roy K. wrote:



Hi Nan,

At the risk of being shot for herecy, I maintain the belief that in the current technological environonment packing everything inside the straightjacket container of a physical file is an anachronism. <snip> So, I would have an attribute in the CF file holding a permanent identifier that is guaranteed to be resolvable from now until eternity into an XML file delivering usage metadata.


I agree with the first statement, because 'everything' is a long list indeed. The world of linked open data agrees too, I expect.

But with respect to *usage* metadata, I wonder. Because the storage of complex data for perpetual access on the web is not a solved problem, and not everyone has access to the entire web, or at least every metadata provider for every file they are working with, all the time. So then you're stuck, until the ship gets back or the DSL line is up or the hosting site finishes its weekly maintenance, or someone goes back and resuscitates some old system that had the metadata years or decades before.

John--
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.




--
*******************************************************
* Nan Galbraith                        (508) 289-2444 *
* Upper Ocean Processes Group            Mail Stop 29 *
* Woods Hole Oceanographic Institution                *
* Woods Hole, MA 02543                                *
*******************************************************
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cgd.ucar.edu/pipermail/cf-metadata/attachments/20110806/223d5f23/attachment-0001.html>
Received on Sat Aug 06 2011 - 01:21:56 BST

This archive was generated by hypermail 2.3.0 : Tue Sep 13 2022 - 23:02:41 BST

⇐ ⇒