Re: vocabs, metadata set, datasets

On Mon, Jan 17, 2011 at 10:06:34AM -0800, Karen Coyle wrote:
> >It works well enough for drawing an analogy, but I wouldn't
> >want to paper over the problem, especially the bit about
> >records being about "one entity (e.g. a book)" -- which is
> >in my opinion simply wrong because a typical catalog record,
> >for example, contains descriptive elements not just about a
> >book, but its author, publisher, etc.
> 
> I think we've gone into the "focus" area again here. A library
> record is designed to provide a complete enough description of a
> book (or piece of music, or map, etc.) to fulfill two functions:
> 
> 1) identification of a resource owned or licensed by a library
> 2) user access to that resource
> 
> The inclusion of related things like authors, places of publication,
> etc. are all with the focus of the thing being cataloged, not as a
> way to describe those related things. Authors are described in their
> own records (name authority records), but only the identifier for
> the author record is included in the bibliographic record. (It just
> so happens that the identifier used today is a text string that
> looks a whole lot like a name.) Thus FRBR and FRAD as separate views
> of bibliographic data.
> 
> The *wholeness* of the record is important because it represents a
> description of the thing that is a complete description. While
> individual statements may be usable in other contexts, the library
> function of bibliographic description will always require a
> particular set of statements (at a minimum). I believe we will
> continue to call this set of statements a "record" for a fairly long
> time.

I agree with you on this, and I think it is important enough
that we should say it clearly like this -- somewhere... maybe
in an entry on "Records".  But it should perhaps also be
acknowledged that the Linked Data space does not have a
universally agreed way to express a bounded record.  This is
a gap that DCAM, for example, tried to address.  The RDF
community has spoken of "named graphs" for a number of years,
but the term is still ambiguous (does it refer to quoted
graphs? graph literals? URIs for graphs?  graph stores?).
Standards-based support for provenance in Linked Data will
depend on what is now referred to as "support for multiple
graphs and graph stores".  For the glossary, we should avoid
going into this sort of detail but this is such an important
difference in paradigm that we should try to characterize
the issues in a few sentences somewhere.

Tom

-- 
Tom Baker <tbaker@tbaker.de>

Received on Monday, 17 January 2011 20:24:44 UTC