W3C home > Mailing lists > Public > www-rdf-interest@w3.org > April 2002

Re: Explicit Disambiguation Via RDF bNodes, more Process

From: Graham Klyne <GK@ninebynine.org>
Date: Sat, 27 Apr 2002 09:58:41 +0100
Message-Id: <>
To: Sandro Hawke <sandro@w3.org>
Cc: www-rdf-interest@w3.org
At 01:08 PM 4/26/02 -0400, Sandro Hawke wrote:
>So take your pick: (1) use this approach, (2) allow some messy merged
>graphs, or (3) achieve consensus.  (or find a better approach.)
>Personally, I'd like (3) but I don't know how to do it.  Maybe when
>people start actually merging graphs, there will be enough social
>pressure on whoever looks the messiest to get them to shape up and

I think that last bit is happening in pockets.  I've found that playing 
with the RDFWeb/RDFWho [1] and related vocabulary [2] (across a small range 
of different applications) has helped to clarify some issues in my mind.  A 
useful technique seems to be to leave abstractions like people un-named or 
named with place-holder URIs, and describe them with uniquely identifying 
properties;  e.g.

     genid:GK a wn:Person ; foaf:mbox <GK@ninebynine.org> .

which seems a little bit like your uname approach without the extra vocabulary.

But the messiness will persist, I think.  Lot's of early work seems to use 
URIs in descriptions in ways that raise the what-is-referenced problem.  I 
think we'll want to live with that even as we evolve better ways to reduce 
the messiness for new data.  Loading additional interpretation birden on 
the properties used might help; e.g.

     <http://www.w3.org/> messy:director "TimBL" .

might use messy:director to mean "the web page whose URI labels the subject 
of this property describes an organization that has a director who is a 
person sometimes referenced by the string that labels the object of the 
property".  Some inference rules might allow us to turn this into a less 
messy description:

     [ a tidier:Organization : tidier:homePage <http://www.w3.org/> ]
     [ a tidier:Person ; tidier:nickname "TimBL" ] .

which is more like the kind of approach suggested above.


[1] http://rdfweb.org/

[2] http://xmlns.com/foaf/0.1/

Graham Klyne
Received on Saturday, 27 April 2002 05:30:37 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:44:35 UTC