[eml-dev] EML use, and associatedParty

[Matthew Jones]

Hi Terry,

The metadata exercise that you describe sounds like it was very 
profitable.  I'd love to learn more about your experiences.  The issues 
you bring up regarding how to handle cross-referencing among different 
extensions of the eml-party module are things we've encountered as well 
when supporting the references tag in Morpho.  I think it is an area we 
should try to strengthen in the spec by making it less ambiguous how to 
interpret the role attribute when using references among associatedParty 
elements.  Your second interpretation is more correct, that 'role' is an 
extension of the ResponsibleParty type, and so:

 > A sequence of (
 >     A choice of (
 >         [contact info.]
 >         OR
 >         [references]
 >     )
 >     [role]
 > )

is right.  Consequently, if you have an associatedParty with a 
'references' field, it must also have a role element, and it is this 
role element that is relevant to that instance of associatedParty.  For 
example:

<associatedParty id="p1">
   <individualName><surName>Jones</surName></individualName>
   <role>Surveyor</role>
</associatedParty>

<associatedParty>
   <references>p1</references>
   <role>Slacker</role>
</associatedParty>

So, the second associatedParty in the example has role 'Slacker', and 
the role from the first associatedParty isn't even contained within the 
scope of the referenced content, although this isn't obvious from the 
XML instance -- you really have to look at the schema to see how the 
inheritance works.  Thus its a bit ambiguous and should be clarified.

The web-based EML editor you developed would probably be of general 
interest to the community.  Is there a demonstration site?  Would you be 
willing to open-source license it for others to use and modify?  Could 
we consider shipping it with EML as we do several other utilities, such 
as the EML Parser that we developed?

We've developed some translators with our collaborators between EML and 
the NBII BDP standards, and these are being used to gateway content from 
the KNB into the NBII metadata clearinghouse.  I suspect the EPA 
Exchange Network support you've developed would also be of general 
interest.  Would you also consider whether that is something that might 
ship with EML?

Thanks for the note.  Looking forward to future discussions.

Cheers,
Matt

Terry Brown wrote:
> Hi,
> 
> [please forward this as necessary]
> 
> I'm Terry Brown, a researcher at U of MN.  My department recently 
> headed up a metadata / data collection and archiving exercise for 
> five $6M EPA STAR projects, each collecting diverse coastal data 
> sets.  http://eagle.nrri.umn.edu/ gives some background, but the 
> short story is that we picked EML to describe and document the 
> data sets, which covered the gamut from genomics lab. results through 
> water sampling and fish netting to remote sensing analysis.
> 
> We put together an entirely server side web based EML editor, and 
> also developed a data flow for the EPA's Exchange Network 
> (http://www.exchangenetwork.net/) which allows transfer of EML+data 
> to the EPA via their CDX protocol.
> 
> Working with EML and training biologists to work with EML has been 
> an interesting process, and we've been meaning to collect our
> thoughts on some of the issues we've encountered and talk with you 
> about them.  We haven't done this yet, but I wanted to give you an 
> idea about the scope of our work with EML before diving into 
> specifics.  Would you be interested in such a discussion when we've 
> cataloged out thoughts?
> 
> That said, the specific that I'm wondering if you have any guidance 
> on at the moment is to do with associatedParty.  Our system 
> currently has 218 EML files (spread amongst the 5 major projects, 
> describing various subprojects in each).  The same individual occurs 
> both in primary roles such as creator, metadataProvider, and 
> associatedParty type roles, this happens a lot.
> 
> I understand why metadataProvider can't cross reference 
> associatedParty... the trouble is I'm not really sure I believe 
> associatedParty can cross reference metadataProvider, either.
> 
> Our system expands cross references during export, and its current 
> behavior is to replace any element that has a non-empty reference 
> sub-element with the element to which the xref points.  When you try 
> and do this for an associatedParty pointing to a metadataProvider, 
> the role is lost.  Actually the system also issues warnings about 
> cross references between dissimilar entities and a cross reference 
> shadowing data (the role).
> 
> This issue wouldn't arise if associatedParty was a more generic 
> element that had a role element and a responsibleParty element as 
> members, rather than being a descendant of responsibleParty.
> 
> Perhaps, as is so often the case, I've answered my own question by 
> trying to spell it out.  I've been thanking of an associatedParty as
> 
> A choice of (
>     A sequence of (
>         [contact info.]
>         [role]
>     )
> OR
>     [reference]
> )
> 
> but should I be thinking of it as:
> 
> A sequence of (
>     A choice of (
>         [contact info.]
>         OR
>         [reference]
>     )
>     [role]
> )
> 
> then at least it would make sense that an xref from associatedParty 
> to metadataProvider replace only the responsibleParty part and 
> retain the role... but it still leaves an inconsistency when there's 
> an xref from one associatedParty to another... which role wins?
> 
> I'd be interested in any thoughts you have on this case and whether 
> or not you know of any similar cases elsewhere in the system?
> 
> Thank you,
> 
> -----------------------------------------------------------------------
> Terry Brown               Natural Resources Research Institute
> Research Associate        5013 Miller Trunk Highway
> tbrown at nrri.umn.edu       University of Minnesota, Duluth
> Ph. 218 720 4345          Duluth, Minnesota 55811
> Fax 218 720 4328          http://beaver.nrri.umn.edu/~tbrown/
> 
> _______________________________________________
> Eml-dev mailing list
> Eml-dev at ecoinformatics.org
> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/eml-dev
> 
> 

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Matthew B. Jones
jones at nceas.ucsb.edu                Ph: 541-888-2581 ext. 287
National Center for Ecological Analysis and Synthesis (NCEAS)
UC Santa Barbara
http://www.nceas.ucsb.edu/ecoinformatics
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~