W3C home > Mailing lists > Public > public-xg-lld@w3.org > January 2011

Re: vocabs, metadata set, datasets

From: <gordon@gordondunsire.com>
Date: Fri, 21 Jan 2011 09:46:07 +0000 (GMT)
To: Karen Coyle <kcoyle@kcoyle.net>, Thomas Baker <tbaker@tbaker.de>
Cc: Mark van Assem <mark@cs.vu.nl>, public-xg-lld@w3.org
Message-ID: <1459455028.97917.1295603168109.JavaMail.open-xchange@oxltgw15.schlund.de>

This is touching on two other issues we have been discussing: granularity, and
data packaging.
Take a simple "record" as an aggregation of some triples:
:R :p1 "text".
:R :p2 :Z.
:R :p3 "other text".
Librarians are used to seeing this record with Z substituted by an appropriate
literal (usually something akin to the value of the corresponding
:R :p1 "text".
:R :p2 <:Z skos:prefLabel> "label". // this touches on the data packaging issue
(btw, is there a syntax to show a triple chain like this?)

:R :p3 "other text".
So the subject of the record is apparently exclusively the resource R. Note that
this pertains to any level of granularity of the record (isbd:Resource,
frbrer:Work, frbrer:Expression, etc.)
But in RDF:
If :p2 owl:InverseOf :p4. // e.g. p2 has label "hasAuthor", p4 has label
Then :Z :p4 :R.
And we have an inferred triple with the subject Z (typically a person,
organization, subject, etc.). Which can be aggregated into an authority record:
:Z skos:prefLabel "label".
:Z :p4 :R.
And the subject of the record is apparently exclusively the entity Z.
So there are different points-of-view, and it is probably always possible to
constrain a "record" to triples with one, and only one, subject. Does RDF let us
have the cake, and eat it?


On 21 January 2011 at 03:15 Thomas Baker <tbaker@tbaker.de> wrote:

> On Thu, Jan 20, 2011 at 02:59:23PM -0800, Karen Coyle wrote:
> > >I see the argument but still see a problem with the notion of
> > >a record being "about" something.  Looking at a bibliographic
> > >record through traditional bibliographic glasses, one does see
> > >a set of data elements "about" a book.  Looking at that same
> > >record through RDF glasses, however, one may see statements
> > >"about" several different things -- the book, its author,
> > >the publisher, etc.
> >
> > As I said before, I don't see it this way. Even though RDF is much
> > more flexible than the old record model, it has the concept of
> > "subject" -- the subject of the statements, in my mind, defines the
> > "about".
> Yes, agreed.
> >          Library data contains things like author identification as
> > an object, but the author is a subject only in the name authority
> > record (or foaf Person description set). I really don't see
> > statements about authors, publishers, etc. in a library catalog
> > record for a book. All of the data there should be with the book as
> > the subject.
> I'm not sure I follow you...  Are you saying that all of the
> information in a library record can always be expressed with
> triples with the same subject?  Even, for example, if FRBR
> distinctions are introduced?
> > In other words, what Jeff said.
> Jeff introduced the unimarc notion of "primary entity" which,
> if I correctly infer, nicely captures the notion that a record
> is focused on a particular resource without implying that all
> of the information in that record may be expressed directly
> as an attribute of the primary entity.
> Tom
> --
> Tom Baker <tbaker@tbaker.de>
Received on Friday, 21 January 2011 09:46:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:35:56 UTC