W3C home > Mailing lists > Public > semantic-web@w3.org > November 2007

Re: Linked Data on qdos.com

From: Frederick Giasson <fred@fgiasson.com>
Date: Thu, 29 Nov 2007 11:54:06 -0500
Message-id: <474EEEAE.3070303@fgiasson.com>
To: Eyal Oren <eoren@few.vu.nl>
Cc: Steve Harris <steve.harris@garlik.com>, Chris Bizer <chris@bizer.de>, Semantic Web Interest Group <semantic-web@w3.org>

Hi Eyal,
> On 11/29/07/11/07 10:49 -0500, Frederick Giasson wrote:
>>> The links to eg. wikipedia use the property 
>>> <http://qdos.com/schema#wikipedia>, is there something we can 
>>> usefully make those properties a subPropertyOf? I would have liked 
>>> to use the FOAF site membership properties, but they're unhelpful 
>>> for the particular data we're trying to handle.
> [..]
>> Also, I would suggest that you use rdfs:domain and rdfs:range to 
>> restrict the use of these properties. Otherwise, with the current 
>> state of the ontology; I can use any of these properties to relate 
>> any resource with any resource :)
> Just as a reminder, according to standard RDF(S) semantics, defining 
> rdfs:domain and rdfs:range do not *restrict* the usage of the property 
> at all. Instead, defined domains and ranges would lead to deriving the 
> rdfs:Class of resources that are described using those properties.
>
I knew someone would popup with that :)

>> so, currently I can say: <foaf:Person> <q:ebay> <foaf:Person>
> You could always say this, with or without a domain/range description.

Well sure; I can always describe the fact that my soul is blue (even if 
a soul can't really has a color). But does loosing people in these 
matters helps the development of the semantic web? Don't think so. I 
mean; yes you are right; but it only get things more obscures when the 
result is somewhat the same.


Take care,



Fred
Received on Thursday, 29 November 2007 16:54:59 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:19 GMT