RE: VIAF contributor model

Very helpful, yes.  And I also would prefer to see us break with past practice of artifically separating the data used as a subject from the data used as a name in whatever other role, and cover the relationship information outside of how we identify the entity.  I think that offers a lot more future flexibility.  And I think that's also what you are saying below. -  Barbara

-----Original Message-----
From: Dan Brickley [mailto:danbri@danbri.org] 
Sent: Tuesday, November 09, 2010 5:31 PM
To: Tillett, Barbara
Cc: Young,Jeff (OR); public-lld
Subject: Re: VIAF contributor model

On Tue, Nov 9, 2010 at 11:11 PM, Tillett, Barbara <btil@loc.gov> wrote:
> I still remain concerned that by using SKOS for names of persons and 
> corporate bodies, there is either an explicit or implied "is the 
> subject of"  relationship going on for the person/corporate body being 
> described with respect to some work.  Am I wrong? - Barbara Tillett

If SKOS feels awkward here, would it help if we added a class foaf:NameDetail, and a property foaf:nameInfo that could link a 'NameDetail' from some Person, Organization, Group or Project? Of course then the name info would need some properties, ... I'd suggest using foaf:name for most simplistic form. I would like to do this
*anyway* for FOAF, so that we can do things like include pronunciation information. If it would be useful here, it can be done more quickly.

FWIW I don't think SKOS is necessarily wrong here, although it wouldn't be my first choice. SKOS concept descriptions are designed to be used when talking about the subject of some work, ... but I think they're formally 'free standing', and can be related to entities of all kinds by a variety of relationship types. Names are funny things, and since FOAF doesn't currently represent names-as-things, SKOS could be used to fill that gap.


So I don't think there's a problem with an implicit 'is the subject of' relationship, even though the concept scheme might have been designed by someone with such relationships very much in mind. For example, a document might have as a subject the thing identified as <http://id.loc.gov/authorities/sh2001001711#concept>; but equally a person might stand in a 'highlySkilledAt' or 'seeksTuitionIn'
relationship to that self-same thing.  The VoiceXML markup language, in this example. In either case the relationship is stated explicitly and is available in the RDF graph for querying/matching etc. None of the formal bits of the SKOS spec imply these missing 'subject' links; if they're not stated explicitly, they're not there.

All this has some potential for bringing a common approach to how we talk about the subject of documents, and the skills and information needs of their users. It is also something of a break from the tradition that subject vocabularies are organised in terms of an implicit 'subject of the document'. However it doesn't stop people organising their schemes that way; rather it just asks that those implicit links are represented explicitly, so that they can sit alongside new kinds of link without confusion.

Hope that makes some sense and I interpreted your question as intended,

cheers,

Dan

Received on Tuesday, 9 November 2010 22:45:44 UTC