KMLM List

["Brian Tingle" <Brian.Tingle@xxxxxxxx>]

Thanks Grace, that's great.  I don't think physically representing these
relationships in the id is a good idea, my interest is more at what
level they can be assigned (and then also mapping them for
disambiguation and reporting).

 

Will the work of this group be applicable only to extant institutions?
I'm thinking of the case of Encoded Archival Context where we will be
creating  records for historic institutions named in our EAD corpus that
may no longer exist.

 

Also, is there any overlap with location identifiers?  I'm thinking of
the use case of a university portal that would have a directory
including campuses, departments and programs, medical centers,
multi-campus research units, arbitrary sub-institutions, that would also
would want to list certain facilities or locations such as
entertainment/sports venues, natural reserves, satellite facilities, and
maybe county agricultural offices that are not necessarily
organizational subunits?  But not at the level of detail that facilities
or media services would need.  (But being able to map to an
institutional identifier from those crazy MARC location codes in the
eight fifty whatever where the call number is would be sweet.)

 

While I'm at it, how does this work relate to ISNI?  ISNI identified
parties "can be natural persons (a human being like an book author),
legal entities (like a Record Label) or even fictional characters (like
Peter Pan)."[1]  If I squint, an institution looks like it is somewhere
between a legal entity (all of the University of California is the same
legal entity) and a fictional character.

 

[1] http://www.isni.org/

 

I suspect intuitional identifiers could also be very helpful in some of
the disambiguation use cases in grants management and technology
transfer as well.

 

Sorry for so many questions, none of them are urgent.  I'm excited to
learn more as your initiative progresses.

 

-- Brian 

 

From: Grace Agnew [mailto:gagnew@xxxxxxxxxxxxxxx] ;
Sent: Thursday, February 18, 2010 6:25 AM
To: Brian Tingle
Cc: i2info@xxxxxxxxxxxxx
Subject: Re: [i2info] i2 question

 

Brian,

These are good questions, and I think we are still considering our
approach to relationships.  We received some comments along these lines
to our survey of library workflows, so your situation is not unique.  In
one of our discussions, it was proposed that the need for a ID is very
contextual, as you say, based on the organizational business need.  A
public library may be a single entity for subscriptions to electronic
databases but may purchase monographs at the branch level, so both
entities in the hierarchy would need identifiers, and sometimes the
branch would have a separate identifier and other times be subsumed as a
subunit under the library system identifier.  We provisionally decided
that any organization could apply at the whole or part organizational
level for as many identifiers as needed by their workflow and that
relationship would be built into the I2 structure to reflect the
relationships among the entities identified.  We haven't decided how to
reflect the relationships, but most likely via registry metadata, as the
easiest strategy requiring the least infrastructure investment for
implementing I2.

It is important to note the distinction between reflecting relationships
among entities and among identifiers.  We expect the identifiers to be
"dumb" unique strings, and assigned at point of need, and therefore not
related to each other.  We do not  propose, for example, the use of
prefixes to concatenate identifiers belonging to related entities
together. We also note that relationships can be quite complex,
particularly as organizations work more collaboratively in consortia to
do things such as purchase electronic resources.  We expect to enable
entities to relate to one another to support the business model that
required the identifier, but we do not plan to assign identifiers that
inherently relate to other identifiers.

So the short answer is that we are very aware of this need--both from
experiences in our own settings and from the feedback we received during
community scenario information gathering.  We plan to address this
situation and are still in the discussion phase about it.  We'll post
our discussions and provisional decisions about this and other issues at
the website and solicit community feedback.  Any ideas/suggested
solutions are extremely welcome!

Grace Agnew
I2 Working Group Co-Chair

 Brian Tingle wrote: 

Hi,

 

I was wondering if the I2 work is going to address the use case of
institutions with arbitrary internal hierarchical (or even
non-hierarchical) relationships.  Or, trying to ask the same question
from a different direction, what is the level of organizational subunit
that can be assigned an identifier, and will the work address the issue
of organizational relationships between institutions?

 

I'm working on three or four different projects where this issue comes
up.  I often observe institutions where, in context A they may represent
their identity at a  different level of organizational subunit
granularity than in context B.  Is this an issue you are grappling with
/ are going to solve for me? J

 

-- Brian

 

 

-- 



Grace Agnew

Associate University Librarian for Digital Library Systems

Rutgers University Libraries

47 Davidson Road

Piscataway, NJ 08854

Phone: 732-445-5908

Fax: 732-445-5888