W3C home > Mailing lists > Public > public-xg-lld@w3.org > October 2010

RE: (Library) Linked Data definition?

From: Young,Jeff (OR) <jyoung@oclc.org>
Date: Thu, 21 Oct 2010 21:02:03 -0400
Message-ID: <52E301F960B30049ADEFBCCF1CCAEF590A23C6D0@OAEXCH4SERVER.oa.oclc.org>
To: "Ford, Kevin" <kefo@loc.gov>, "public-xg-lld" <public-xg-lld@w3.org>
Kevin,

In one sense, "Linked Data" is simply a mechanism to identify things
that aren't Web documents (aka "real world objects"). Past attempts to
do this with URNs and "info" URIs weren't productive, so the W3C devised
an HTTP URI solution using either 303 (See Other) redirects or hash
URIs. Either way, the HTTP protocol leads naturally to a different URI
where information *about* the identified "real world object" can be
obtained. Content-negotiation for RDF or HTML representations can be
used to deliver this information for both human and machine consumption.
Details of the mechanism can be found here:

http://www.w3.org/TR/cooluris/
 
The ability to unambiguously identify "real world objects" creates vast
opportunities for linking/equating things, but also raises deep
ontological questions. Those questions need to be asked in the context
of use cases which themselves are often vague or poorly understood.
Assuming agreeable models can be derived from the use cases, there is
still the problem of communicating and automating the models. 

I like OWL.

Jeff

> -----Original Message-----
> From: public-xg-lld-request@w3.org [mailto:public-xg-lld-
> request@w3.org] On Behalf Of Ford, Kevin
> Sent: Thursday, October 21, 2010 4:07 PM
> To: 'public-xg-lld'
> Subject: (Library) Linked Data definition?
> 
> Dear All,
> 
> I was looking over the Use Cases and I couldn't readily see how a few
> of them might be considered (Library) Linked Data use cases.  I began
> wondering if it was because the Use Case was unclear or if it was
> because I had a different (or also unclear) notion of Linked Data.
For
> example, did the use case clearly articulate who the users would be
and
> how the those users would interact with a given system?  If those
> conditions were met, the reviewer would be able to evaluate how
> conducive the Use Case was to the Linked Data model and, therefore,
> what needed to be developed (ontology, communications protocol) to
> realize the Use Case (if not already realized).
> 
> But, whether the Use Case was clear or not, I noted that there does
not
> seem to be a working definition of 'Linked Data' that one could use,
to
> determine on a case-by-case basis whether a given use case is indeed a
> Linked Data Use Case.  I came to this group late, but I was unable to
> locate any formal definition on the Wiki (I did find Linked Data
> resources and terminology, and Library standards and linked data,
etc).
> 
> Is there such a working definition, and if not, wouldn't it be
> essential to develop a definition, as painful as that process may be?
> This might help in evaluating the Use Cases.
> 
> Rgds,
> Kevin
> 
> 
Received on Friday, 22 October 2010 01:10:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 22 October 2010 01:10:52 GMT