Toward EML2.0

Tue Aug 20 14:25:51 PDT 2002

I should respond as i dont think corinna's and my objections about
identifiers really dealt with this issue. Identifiers as they are being used
in Beta9 are not necessarily identifiers of an object (in this case a
person) - they are identifiers of a fragment of EML that describes that
object. 

Matt's example in fact potentially raises some of these issues. According to
the rules in:
http://knb.ecoinformatics.org/software/eml/eml20docs/eml-docbook.html#reusab
leContent,
I cannot use the same identifier for two ResponsibleParty segments that
describe Nancy Grimm if one of those uses her Biology departnemt affiliation
and another uses her CES affiliation (or if one required more detail than
another such as including a middle initial that wasnt used in one
publication). The reason for this is that the content would be different,
and thus cannot have the same identifier - it would raise a key violation.
Similarly, because i chose to use a system-scoped identifier, I also cannot
use the same identifier for Nancy once i have updated her information in
that ldap, because it would conflict with the information I had provided in
previous metadata documents in which her old phone number appeared. Although
no key violation would be raised between the documents, I'm nevertheless
violating the rules as stated above - the identifier would not be
referencing the same content, despite the fact that its clearly the same
person we're talking about. Now if this is ok to do, then I have
misunderstood our debates over this point, because it seemed clear to me
that the NCEAS opinion about identifiers was that if I identified a chunk of
information as "x", then its content should not vary in any way from any
other past, present or future chunk that is also identified as "x",
regardless of whether both chunks were attempting to describe the same
thing(or in this case person). Correct me if I'm wrong. 

My problem with providing identifiers was, and is, that they lead to
problems if you are going to expect them to have any meaning outside a
narrowly bounded context. In this case, EML has defined rules about the
relationship between an identifier and its content that the extenal system
(CAP Personnel database) being cited does not live up to (assuming that it
was ok for that personnel directory to update Nancy's phone number without
assigning a new user_id).

Peter McCartney (peter.mccartney at asu.edu)
Center for Environmental Studies
Arizona State University
480-965-6791 

-----Original Message-----
From: Matt Jones [mailto:jones at nceas.ucsb.edu]
Sent: Tuesday, August 20, 2002 9:02 AM
To: Karen Baker
Cc: eml-dev at ecoinformatics.org; im at lternet.edu
Subject: Re: Toward EML2.0

Hi Karen,

Thanks for the note.  A few comments...

We have recognized the need for an English language specification that 
accompanies the schemas and describes aspects of EML that are not 
described in the schemas themselves.  The current beta 9 spec has an 
initial attempt at providing such a spec, and can be found at:
http://knb.ecoinformatics.org/software/eml/eml20docs/eml-docbook.html
It needs work -- any contributions to writing this document would be 
greatly appreciated, as well as suggestions as to how the document 
should be structured.  Right now, we are automatically generating the 
technical aspects of the document directly from the schemas, and we are 
generating the rest from Docbook sources that we've written.

I completely agree about the party module.  A long while ago (at the 
first EML workshop in Albuquerque), and in earlier versions of EML, I 
had argued for and implemented an identifying attribute for parties 
(e.g., a username or userid field) that was specific within some 
authentication system so that parties in EML documents could be tied 
directly to authentication databases like the ecoinfo/nceas/lter ldap 
server.  Peter and Corinna and others argued strongly that such an ID 
was completely unneeded, and so it was removed several versions back. 
Under the new revision of beta9, however, we have re-instituted the 
concept of an "id" in party as part of the referencing scheme.  So now, 
presumably, a party could be listed with an id and a reference to an 
authentication system such as the ecoinfo ldap server, and it would be 
unambiguous.  So, I think the current system in Beta9 does in fact allow 
strong ties to personnel databases. Do you agree?  Here's an example 
snippet that would be eml beta9 compliant:

<creator id="uid=jones,o=NCEAS,dc=ecoinformatics,dc=org" 
system="ldap://ldap.ecoinformatics.org/dc=ecoinformatics,dc=org" 
scope="system">
   <individualName><surName>Jones</surName></individualName>
</creator>

Regards,
Matt

-- 
*******************************************************************
Matt Jones                                    jones at nceas.ucsb.edu
http://www.nceas.ucsb.edu/    Fax: 425-920-2439   Ph: 907-789-0496
National Center for Ecological Analysis and Synthesis (NCEAS)

Interested in ecological informatics? http://www.ecoinformatics.org
*******************************************************************

_______________________________________________
eml-dev mailing list
eml-dev at ecoinformatics.org
http://www.ecoinformatics.org/mailman/listinfo/eml-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/eml-dev/attachments/20020820/c85a9cc4/attachment.htm