W3C home > Mailing lists > Public > www-rdf-dspace@w3.org > May 2004

RE: Overview of design decisions made creating stylesheet and schema for Artstor data

From: MacKenzie Smith <kenzie@MIT.EDU>
Date: Thu, 20 May 2004 14:11:33 -0400
Message-Id: <5.2.1.1.2.20040520134819.01c97e90@hesiod>
To: "Butler, Mark" <mark-h.butler@hp.com>
Cc: www-rdf-dspace@w3.org

Hi Mark,

Sorry to chime in late, but I'm interested to know where this ended up.

David didn't say that he objected to calling something a "creator", he said 
that creators can have different roles in the creation process (e.g. be an 
author in one case and an editor in another case) and that these roles show 
up as attributes of the creator in specific contexts. One wants to be able 
to filter on "T.S. Eliot as an author" (creator.role="author" and NOT 
creator.role="editor"). If role is a repeatable property of a creator then 
T.S. Eliot has two associated roles (true) and multiple works associated 
with him (true) but the relationship between the specific role and the 
specific work is lost.

So the issue with this example is that the creator entity, if 
normalized/unified/merged, could end up with multiple 'role' properties and 
no way to tell which of them applies to the work being described... in 
other words, in this case one particular attribute (role) is very 
context-specific so that normalizing/unifying/merging instances of the 
creator metadata would create a problem.

What we would typically do in this case is normalize/unify/merge the bits 
that are invariant across instances (i.e. first and last name, birth and 
death dates, etc.), but have the variant attributes (e.g. role) be related 
first class objects.

MacKenzie

At 10:31 AM 5/5/2004 +0100, Butler, Mark wrote:
>I think I understand now. You don't like the objects being called
>"Creators". In the end, we didn't do that - they are called "Entity". To
>clarify, here is an example from the artstor dataset.
[snip]
>  <vra:creator>
>     <vra:Entity
>rdf:about="http://web.mit.edu/simile/www/metadata/artstor/Creator#van_dyck,_
>anthony,sir,1599-1641">
>   <rdfs:label>Van Dyck, Anthony,Sir,1599-1641</rdfs:label>
>   <rdf:type
>rdf:resource="http://web.mit.edu/simile/www/2003/10/person#Person"/>
>   <vc:Family>Van Dyck</vc:Family>
>   <vc:Given>Anthony</vc:Given>
>   <vc:FN>Anthony Van Dyck</vc:FN>
>     </vra:Entity>
>  </vra:creator>
Received on Thursday, 20 May 2004 22:03:04 EDT

This archive was generated by hypermail pre-2.1.9 : Thursday, 20 May 2004 22:03:07 EDT