[Esip-documentation] Minutes / handling of ACDD guidance

[John Graybeal]

I'm not sure what it means to say "the ACDD documentation would be amended", or whether that's a final decision. But I have a proposal intended to address this question, and avoid other process debates peeking their heads up....

As this thread every so often gets questions about the meaning/application of ACDD, indicating the specification is being actively used (yay) and not every answer is self-explanatory (shocking :->).  I propose we create a an ACDD guidance page, which could briefly explain best practices and understandings as they are derived in the emails.

In the present example, the page could include the facts that: 
* the lat/lon/vertical resolution variables only should be used if there is a grid of observations in that axis
* tools typically construe a value without units to be the same units as the units given by geospatial_lat/lon/vertical_units
* values following the form "number units" should use UDUNITS-compatible units if they are to be understood by tools,
and possibly, depending on how the conversation unfolds, 
* lat/lon resolutions should only be given for georectified grids (list of acceptable georectified units would be helpful here).

(If the idea that resolution attributes should not include units takes hold, that would require an update to the spec. We don't seem to be heading that direction though.)

Perhaps we could even agree that a small help icon in the main document's attributes table could be added whenever a guidance topic is added, without changing the ACDD specification version.

Thoughts?

John



On Feb 26, 2015, at 11:37, Ted Habermann via Esip-documentation <esip-documentation at lists.esipfed.org> wrote:

> Hello all,
> 
> These are the minutes for today's meeting. I will get them on the Wiki/commons ASAP...
> 
> Ted
> 
> Documentation Cluster Minutes - February 2015
> 
> The Documentation cluster meeting was delayed a week then held two hours early (sorry about the confusion caused by Ted).
> Nevertheless, we had an interesting discussion.
> 
> 1. Resolution and units
> There was a long email thread yesterday started by Aaron Sweeny about how horizontal resolutions were described in netCDF and in the ACDD documentation. There was confusion caused by the recommendation that resolution strings in ACDD have the form “number units”. The current transform translates this into invalid ISO because the ISO value includes the entire string and it is expected to be a numeric value (needs to be checked). There is also a situation in an NODC template where, for point data, the resolution is recommended to be the string “point”. This also causes problems related to the type of the result in ISO.
> 
> It was decided that the ACDD documentation would be amended to include that fact that, if the resolution is given without units, it is assumed to be the same as the unit given by geospatial_lat/lon_units.
> 
> 2. A way forward
> 
> Aaron is also interested in how Sea Data Net attributes can be translated from netCDF to ISO. The only recommendation for a way forward involved branching the ncISO xsl (in gitHub at https://github.com/Unidata/threddsIso/tree/master/src/main/resources/xsl/nciso) to include those attributes as input. This branch could then be described on the ACDD pages in the ESIP wiki. The SeaDataNet metadata page is at http://www.seadatanet.org/Standards-Software/Metadata-formats.
> 
> 3. Metadata Recommendations and Dialects
> 
> Ted reported briefly on the work he has been doing with tools for comparing metadata recommendations and dialects. Early versions of these tools are at:
> 
> http://ec2-54-149-127-50.us-west-2.compute.amazonaws.com:443/drTool/
> 
> http://ec2-54-149-127-50.us-west-2.compute.amazonaws.com:443/recTool/ and
> 
> http://ec2-54-149-127-50.us-west-2.compute.amazonaws.com:443/dialectTool/
> 
> Maybe next month we will have more time for a complete description of this work...
> 4. Linked Data Dataset Description
> 
> Ted reported on some recent metadata developments in the W3C Health Care and Life Sciences Working Group. This group is working on a profile for dataset descriptions that brings together many commonly used linked data namespaces. The Editors Working Draft of this profile is at http://www.w3.org/2001/sw/hcls/notes/hcls-dataset/. Ted is working on crosswalks from that profile to metadata dialects that are commonly used in the Earth sciences.
> 
> <SignatureSm.png>
> 
> _______________________________________________
> Esip-documentation mailing list
> Esip-documentation at lists.esipfed.org
> http://lists.deltaforce.net/mailman/listinfo/esip-documentation

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.deltaforce.net/pipermail/esip-documentation/attachments/20150227/5644d513/attachment-0001.html>