[eml-dev] [Fwd: [Bug 2512] New: - require text content in elements to be non-empty]

Matt Jones jones at nceas.ucsb.edu
Wed Aug 16 12:24:08 PDT 2006


Hi Inigo,

That's an intersting field that you chose.  It is intended to explain 
what a missing value code means when more than one is used.  For 
example, sometimes people might use -999 to mean "missing because 
sampling not performed" but maybe -888 might mean "missing because 
organism not found".  In the absence of these sorts of subtleties, all 
missing value codes could be interpreted as just "missing", so I think 
at first glance that this codeExplanation field could be made optional 
without loss of functionality.  I'm surprised its hard to fill in as for 
most people a simple phrase like "Missing value" would be a sufficient 
default if they don't differentiate the different kinds of codes, and if 
they do differentiate then they should know what the code means and 
could fill it in.

Matt

inigo san gil wrote:
> Matt Jones wrote:
> 
>>Thanks, Inigo.
>>
>>I agree we won't prevent all of these issues.  But at least we can 
>>make it explicit that when we make an element required it is because 
>>we want real content there.
>>
>>I also agree that we should discuss common situations where people 
>>have used this trick to avoid providing content and consider revising 
>>the requirements.  The 'precision' field was initially required but 
>>later made optional in EML 2.0.1 for this very reason.  Other fields 
>>may be in the same category.  However, I would note that we already 
>>struggle with automating data import with the existing set of required 
>>information, and so loosening up the requirements will negatively 
>>impact what we can do with the metadata.
>>
>>Do you have any suggested fields to make optional that wouldn't 
>>adversely influence the effectiveness of the metadata?
>>
> 
> 
> i have some issues in the spatialRaster branch, mostly to create a 
> better crosswalk with the other big standard FGDC and related profiles), 
> and I'll be happy to share with you all. I talked to Mark Servilla about 
> some of these, and we arrived a compromise to savage a lot of metadata 
> making one or two "unknown" entries.
> 
> .. but here is one that comes very often:
> 
> eml/dataset/dataTable/attributeList/attribute/missingValueCode/codeExplanation  

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Matt Jones                                   Ph: 907-789-0496
jones at nceas.ucsb.edu                    SIP #: 1-747-626-7082
National Center for Ecological Analysis and Synthesis (NCEAS)
UC Santa Barbara     http://www.nceas.ucsb.edu/ecoinformatics
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


More information about the Eml-dev mailing list