W3C home > Mailing lists > Public > www-archive@w3.org > December 2007

POWDER: reification or what

From: Jeremy Carroll <jjc@hpl.hp.com>
Date: Fri, 14 Dec 2007 21:33:02 +0000
Message-ID: <4762F68E.8080109@hpl.hp.com>
To: www-archive@w3.org, Phil Archer <parcher@icra.org>, "Carroll, Jeremy John" <jeremy.carroll@hp.com>, Dan Brickley <danbri@danbri.org>, "Williams, Stuart (HP Labs, Bristol)" <skw@hp.com>
CC: Ivan Herman <ivan@w3.org>, Bijan Parsia <bparsia@cs.man.ac.uk>

Jeremy Carroll wrote:

> - reification, or what?
>    I dislike reification, and will argue against it, and in favour of 
> the design in the powder-dr WD, with minor mods.

My problem with reification is that the typical use case, like this one, 
requires quoting, and reification does not deliver that, as is made 
clear in the relevant parts of RDF Semantics.

The only compelling argument agsinst it, though, is somewhat abstract, 
and is the Clark Kent vs Superman.

Lois Lane says "Lois loves Superman"
Lois Lane says "Lois hates Clark Kent"

I know that Clark Kent and superman are the same person, so I can 
truthfully, but somewhat disingenuously say:

Lois said that she loves Clark Kent and hates superman.

But if I use quoting the sentence

Lois said "Lois loves Clark Kent"

is false.

In the example in

The description resource is used as the object to which the metadata is 

1  <wdr:DR>
2    <foaf:maker rdf:resource="http://authority.example.org/foaf.rdf#me" />
3    <dcterms:issued>2007-07-02</dcterms:issued>
4    <wdr:validFrom>2008-07-07</wdr:validFrom>
5    <wdr:validUntil>2008-07-07</wdr:validUntil>

6    <wdr:hasScope>
7      <wdr:ResourceSet>
8        <wdr:includeHosts>example.org</wdr:includeHosts>
9      </wdr:ResourceSet>
10   </wdr:hasScope>

11   <wdr:hasDescriptors>
12     <wdr:Descriptors>
13       <ex:property1>value 1</ex:property1>
14       <ex:property2>value 2</ex:property2>
15     </wdr:Descriptors>
16   </wdr:hasDescriptors>

17   <dc:description>Textual information to display to end 
18 </wdr:DR>

If we modify this simply by changing line 1 to

<wdr:DR rdf:about="">

then we are saying that the current document  (i.e. "") is a description 
resource, i.e. a description of a set of resources. The RDF/XML 
representation of that resource is the given document, which corresponds 
to an RDF graph, which is hence named with the URL of the document. 
(From the point of view of the named graph perspective).

The metadata block, lines 2 to 5, is explicitly metadata about this 
description resource, i.e. the current document.

If we want many such description resources in the same document, we 
could have them by giving each a fragment ID, e.g rdf:ID="dr1", 
rdf:ID="dr2" etc.

While this is not as fine grained as the statement by statement metadata 
provided by the reification solution, I think this solution adequately 
addresses requirement 3.1.10 from UCR.

Personally, I would be happier if this was explicitly performed as named 
graphs, but lacking a standard for serializing named graphs, I believe 
the articulation here is adequate.

Received on Friday, 14 December 2007 21:33:44 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:43:17 UTC